diff --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
index 07a8724b6f33d..96d8300a0faf3 100644
--- a/clang/include/clang/Basic/riscv_vector.td
+++ b/clang/include/clang/Basic/riscv_vector.td
@@ -1013,9 +1013,9 @@ let ManualCodegen = [{
 }] in {
   let HasFRMRoundModeOp = true in {
     // 13.2. Vector Single-Width Floating-Point Add/Subtract Instructions
-    defm vfadd  : RVVFloatingBinBuiltinSetRoundingMode;
-    defm vfsub  : RVVFloatingBinBuiltinSetRoundingMode;
-    defm vfrsub : RVVFloatingBinVFBuiltinSetRoundingMode;
+    defm vfadd  : RVVFloatingBinBuiltinSetRoundingMode<HasBF=1>;
+    defm vfsub  : RVVFloatingBinBuiltinSetRoundingMode<HasBF=1>;
+    defm vfrsub : RVVFloatingBinVFBuiltinSetRoundingMode<HasBF=1>;
 
     // 13.3. Vector Widening Floating-Point Add/Subtract Instructions
     // Widening FP add/subtract, 2*SEW = 2*SEW +/- SEW
@@ -1023,14 +1023,14 @@ let ManualCodegen = [{
     defm vfwsub : RVVFloatingWidenOp0BinBuiltinSetRoundingMode;
 
     // 13.4. Vector Single-Width Floating-Point Multiply/Divide Instructions
-    defm vfmul  : RVVFloatingBinBuiltinSetRoundingMode;
+    defm vfmul  : RVVFloatingBinBuiltinSetRoundingMode<HasBF=1>;
     defm vfdiv  : RVVFloatingBinBuiltinSetRoundingMode;
     defm vfrdiv : RVVFloatingBinVFBuiltinSetRoundingMode;
   }
   // 13.2. Vector Single-Width Floating-Point Add/Subtract Instructions
-  defm vfadd  : RVVFloatingBinBuiltinSet;
-  defm vfsub  : RVVFloatingBinBuiltinSet;
-  defm vfrsub : RVVFloatingBinVFBuiltinSet;
+  defm vfadd  : RVVFloatingBinBuiltinSet<HasBF=1>;
+  defm vfsub  : RVVFloatingBinBuiltinSet<HasBF=1>;
+  defm vfrsub : RVVFloatingBinVFBuiltinSet<HasBF=1>;
 
   // 13.3. Vector Widening Floating-Point Add/Subtract Instructions
   // Widening FP add/subtract, 2*SEW = 2*SEW +/- SEW
@@ -1038,7 +1038,7 @@ let ManualCodegen = [{
   defm vfwsub : RVVFloatingWidenOp0BinBuiltinSet;
 
   // 13.4. Vector Single-Width Floating-Point Multiply/Divide Instructions
-  defm vfmul  : RVVFloatingBinBuiltinSet;
+  defm vfmul  : RVVFloatingBinBuiltinSet<HasBF=1>;
   defm vfdiv  : RVVFloatingBinBuiltinSet;
   defm vfrdiv : RVVFloatingBinVFBuiltinSet;
 }
@@ -1065,6 +1065,10 @@ let ManualCodegen = [{
         defm vfwmul : RVVOutOp0Op1BuiltinSet<"vfwmul", "x",
                                              [["vv", "w", "wvvu"],
                                               ["vf", "w", "wveu"]]>;
+      let RequiredFeatures = ["zvfbfa"] in
+        defm vfwmul : RVVOutOp0Op1BuiltinSet<"vfwmul", "y",
+                                             [["vv", "vw", "wvvu"],
+                                              ["vf", "vw", "wveu"]]>;
     }
   }
   // 13.3. Vector Widening Floating-Point Add/Subtract Instructions
@@ -1081,6 +1085,10 @@ let ManualCodegen = [{
       defm vfwmul : RVVOutOp0Op1BuiltinSet<"vfwmul", "x",
                                            [["vv", "w", "wvv"],
                                             ["vf", "w", "wve"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm vfwmul : RVVOutOp0Op1BuiltinSet<"vfwmul", "y",
+                                           [["vv", "vw", "wvv"],
+                                            ["vf", "vw", "wve"]]>;
   }
 }
 }
@@ -1170,6 +1178,8 @@ let ManualCodegen = [{
     defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "fd", [["v", "v", "vvu"]]>;
     let RequiredFeatures = ["zvfh"] in
       defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "x", [["v", "v", "vvu"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "y", [["v", "v", "vvu"]]>;
   }
   // 13.8. Vector Floating-Point Square-Root Instruction
   defm vfsqrt : RVVOutBuiltinSet<"vfsqrt", "fd", [["v", "v", "vv"]]>;
@@ -1180,21 +1190,26 @@ let ManualCodegen = [{
   defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "fd", [["v", "v", "vv"]]>;
   let RequiredFeatures = ["zvfh"] in
     defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "x", [["v", "v", "vv"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm vfrec7 : RVVOutBuiltinSet<"vfrec7", "y", [["v", "v", "vv"]]>;
 }
 
 // 13.9. Vector Floating-Point Reciprocal Square-Root Estimate Instruction
 defm vfrsqrt7 : RVVOutBuiltinSet<"vfrsqrt7", "fd", [["v", "v", "vv"]]>;
 let RequiredFeatures = ["zvfh"] in
   defm vfrsqrt7 : RVVOutBuiltinSet<"vfrsqrt7", "x", [["v", "v", "vv"]]>;
+let RequiredFeatures = ["zvfbfa"] in
+  defm vfrsqrt7 : RVVOutBuiltinSet<"vfrsqrt7", "y", [["v", "v", "vv"]]>;
+
 
 // 13.11. Vector Floating-Point MIN/MAX Instructions
-defm vfmin : RVVFloatingBinBuiltinSet;
-defm vfmax : RVVFloatingBinBuiltinSet;
+defm vfmin : RVVFloatingBinBuiltinSet<HasBF=1>;
+defm vfmax : RVVFloatingBinBuiltinSet<HasBF=1>;
 
 // 13.12. Vector Floating-Point Sign-Injection Instructions
-defm vfsgnj  : RVVFloatingBinBuiltinSet;
-defm vfsgnjn : RVVFloatingBinBuiltinSet;
-defm vfsgnjx : RVVFloatingBinBuiltinSet;
+defm vfsgnj  : RVVFloatingBinBuiltinSet<HasBF=1>;
+defm vfsgnjn : RVVFloatingBinBuiltinSet<HasBF=1>;
+defm vfsgnjx : RVVFloatingBinBuiltinSet<HasBF=1>;
 }
 defm vfneg_v : RVVPseudoVFUnaryBuiltin<"vfsgnjn", "fd">;
 let RequiredFeatures = ["zvfh"] in
@@ -1219,6 +1234,8 @@ let UnMaskedPolicyScheme = HasPassthruOperand in {
 defm vfclass : RVVOp0BuiltinSet<"vfclass", "fd", [["v", "Uv", "Uvv"]]>;
 let RequiredFeatures = ["zvfh"] in
   defm vfclass : RVVOp0BuiltinSet<"vfclass", "x", [["v", "Uv", "Uvv"]]>;
+let RequiredFeatures = ["zvfbfa"] in
+  defm vfclass : RVVOp0BuiltinSet<"vfclass", "y", [["v", "vUv", "Uvv"]]>;
 }
 
 // 13.15. Vector Floating-Point Merge Instruction
@@ -1239,6 +1256,9 @@ let HasMasked = false,
   let RequiredFeatures = ["zvfh"] in
     defm vfmerge : RVVOutOp1BuiltinSet<"vfmerge", "x",
                                        [["vfm", "v", "vvem"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm vfmerge : RVVOutOp1BuiltinSet<"vfmerge", "y",
+                                       [["vfm", "v", "vvem"]]>;
 }
 
 // 13.16. Vector Floating-Point Move Instruction
@@ -1252,6 +1272,9 @@ let HasMasked = false,
   let RequiredFeatures = ["zvfh"] in
     defm vfmv_v : RVVOutBuiltinSet<"vfmv_v_f", "x",
                                    [["f", "v", "ve"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm vfmv_v : RVVOutBuiltinSet<"vfmv_v_f", "y",
+                                   [["f", "v", "ve"]]>;
 }
 
 // 13.17. Single-Width Floating-Point/Integer Type-Convert Instructions
@@ -1287,10 +1310,16 @@ let Log2LMUL = [-3, -2, -1, 0, 1, 2] in {
       defm : RVVConvBuiltinSet<"vfwcvt_f_x_v", "c", [["Fw", "Fwv"]]>;
     }
   }
+  let RequiredFeatures = ["zvfbfa"], OverloadedName = "vfwcvt_f_bf16" in {
+    defm : RVVConvBuiltinSet<"vfwcvt_f_xu_v", "c", [["Yw", "YwUv"]]>;
+    defm : RVVConvBuiltinSet<"vfwcvt_f_x_v", "c", [["Yw", "Ywv"]]>;
+  }
   let OverloadedName = "vfwcvt_f" in {
     defm : RVVConvBuiltinSet<"vfwcvt_f_f_v", "f", [["w", "wv"]]>;
     let RequiredFeatures = ["zvfhmin"] in
       defm : RVVConvBuiltinSet<"vfwcvt_f_f_v", "x", [["w", "wv"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm : RVVConvBuiltinSet<"vfwcvt_f_f_v", "y", [["vw", "wv"]]>;
   }
 }
 
@@ -1300,17 +1329,23 @@ let Log2LMUL = [-3, -2, -1, 0, 1, 2] in {
     defm : RVVConvBuiltinSet<"vfncvt_rtz_xu_f_w", "si", [["Uv", "UvFw"]]>;
     let RequiredFeatures = ["zvfh"] in
       defm : RVVConvBuiltinSet<"vfncvt_rtz_xu_f_w", "c", [["Uv", "UvFw"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm : RVVConvBuiltinSet<"vfncvt_rtz_xu_f_w", "c", [["YwUv", "UvYw"]]>;
   }
   let OverloadedName = "vfncvt_rtz_x" in {
     defm : RVVConvBuiltinSet<"vfncvt_rtz_x_f_w", "si", [["Iv", "IvFw"]]>;
     let RequiredFeatures = ["zvfh"] in
       defm : RVVConvBuiltinSet<"vfncvt_rtz_x_f_w", "c", [["Iv", "IvFw"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm : RVVConvBuiltinSet<"vfncvt_rtz_x_f_w", "c", [["YwIv", "IvYw"]]>;
   }
   let OverloadedName = "vfncvt_rod_f" in {
     defm : RVVConvBuiltinSet<"vfncvt_rod_f_f_w", "f", [["v", "vw"]]>;
     let RequiredFeatures = ["zvfh"] in
       defm : RVVConvBuiltinSet<"vfncvt_rod_f_f_w", "x", [["v", "vw"]]>;
   }
+  let RequiredFeatures = ["zvfbfa"], OverloadedName = "vfncvt_rod_f_bf16" in
+    defm : RVVConvBuiltinSet<"vfncvt_rod_f_f_w", "y", [["v", "vw"]]>;
 }
 
 // Zvfbfmin - Vector convert BF16 to FP32
@@ -1363,11 +1398,15 @@ let ManualCodegen = [{
         defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "si", [["Iv", "IvFwu"]]>;
         let RequiredFeatures = ["zvfh"] in
           defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "c", [["Iv", "IvFwu"]]>;
+        let RequiredFeatures = ["zvfbfa"] in
+          defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "c", [["YwIv", "IvYwu"]]>;
       }
       let OverloadedName = "vfncvt_xu" in {
         defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "si", [["Uv", "UvFwu"]]>;
         let RequiredFeatures = ["zvfh"] in
           defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "c", [["Uv", "UvFwu"]]>;
+        let RequiredFeatures = ["zvfbfa"] in
+          defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "c", [["YwUv", "UvYwu"]]>;
       }
       let OverloadedName = "vfncvt_f" in {
         defm : RVVConvBuiltinSet<"vfncvt_f_x_w", "f", [["v", "vIwu"]]>;
@@ -1382,6 +1421,8 @@ let ManualCodegen = [{
         let RequiredFeatures = ["zvfhmin"] in
           defm : RVVConvBuiltinSet<"vfncvt_f_f_w", "x", [["v", "vwu"]]>;
       }
+      let RequiredFeatures = ["zvfbfa"], OverloadedName = "vfncvt_f_bf16" in
+        defm : RVVConvBuiltinSet<"vfncvt_f_f_w", "y", [["v", "vwu"]]>;
     }
 
     // Zvfbfmin - Vector convert FP32 to BF16
@@ -1430,11 +1471,15 @@ let ManualCodegen = [{
       defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "si", [["Iv", "IvFw"]]>;
       let RequiredFeatures = ["zvfh"] in
         defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "c", [["Iv", "IvFw"]]>;
+      let RequiredFeatures = ["zvfbfa"] in
+        defm : RVVConvBuiltinSet<"vfncvt_x_f_w", "c", [["YwIv", "IvYw"]]>;
     }
     let OverloadedName = "vfncvt_xu" in {
       defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "si", [["Uv", "UvFw"]]>;
       let RequiredFeatures = ["zvfh"] in
         defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "c", [["Uv", "UvFw"]]>;
+      let RequiredFeatures = ["zvfbfa"] in
+        defm : RVVConvBuiltinSet<"vfncvt_xu_f_w", "c", [["YwUv", "UvYw"]]>;
     }
     let OverloadedName = "vfncvt_f" in {
       defm : RVVConvBuiltinSet<"vfncvt_f_x_w", "f", [["v", "vIw"]]>;
@@ -1449,6 +1494,8 @@ let ManualCodegen = [{
       let RequiredFeatures = ["zvfhmin"] in
         defm : RVVConvBuiltinSet<"vfncvt_f_f_w", "x", [["v", "vw"]]>;
     }
+    let RequiredFeatures = ["zvfbfa"], OverloadedName = "vfncvt_f_bf16" in
+      defm : RVVConvBuiltinSet<"vfncvt_f_f_w", "y", [["v", "vw"]]>;
   }
 
   // Zvfbfmin - Vector convert FP32 to BF16
@@ -1578,6 +1625,9 @@ let HasMasked = false, MaskedPolicyScheme = NonePolicy in {
     let RequiredFeatures = ["zvfh"] in
       defm vfmv_f : RVVOp0BuiltinSet<"vfmv_f_s", "x",
                                      [["s", "ve", "ev"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm vfmv_f : RVVOp0BuiltinSet<"vfmv_f_s", "y",
+                                     [["s", "ve", "ev"]]>;
   }
   let OverloadedName = "vfmv_s",
       UnMaskedPolicyScheme = HasPassthruOperand,
@@ -1589,6 +1639,9 @@ let HasMasked = false, MaskedPolicyScheme = NonePolicy in {
       defm vfmv_s : RVVOutBuiltinSet<"vfmv_s_f", "x",
                                      [["f", "v", "ve"],
                                       ["x", "Uv", "UvUe"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm vfmv_s : RVVOutBuiltinSet<"vfmv_s_f", "y",
+                                     [["f", "v", "ve"]]>;
   }
 }
 
@@ -1601,11 +1654,11 @@ defm vslidedown : RVVSlideDownBuiltinSet;
 // 16.3.3. Vector Slide1up Instructions
 let UnMaskedPolicyScheme = HasPassthruOperand in {
 defm vslide1up : RVVSlideOneBuiltinSet;
-defm vfslide1up : RVVFloatingBinVFBuiltinSet;
+defm vfslide1up : RVVFloatingBinVFBuiltinSet<HasBF=1>;
 
 // 16.3.4. Vector Slide1down Instruction
 defm vslide1down : RVVSlideOneBuiltinSet;
-defm vfslide1down : RVVFloatingBinVFBuiltinSet;
+defm vfslide1down : RVVFloatingBinVFBuiltinSet<HasBF=1>;
 
 // 16.4. Vector Register Gather Instructions
 // signed and floating type
diff --git a/clang/include/clang/Basic/riscv_vector_common.td b/clang/include/clang/Basic/riscv_vector_common.td
index 767bcee7b1596..2a2a04998366a 100644
--- a/clang/include/clang/Basic/riscv_vector_common.td
+++ b/clang/include/clang/Basic/riscv_vector_common.td
@@ -470,6 +470,10 @@ let HasMaskedOffOperand = false in {
       defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                     [["vv", "v", "vvvv"],
                                      ["vf", "v", "vvev"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vv", "v", "vvvv"],
+                                     ["vf", "v", "vvev"]]>;
   }
   multiclass RVVFloatingTerBuiltinSetRoundingMode {
     defm "" : RVVOutOp1BuiltinSet<NAME, "fd",
@@ -479,6 +483,10 @@ let HasMaskedOffOperand = false in {
       defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                     [["vv", "v", "vvvvu"],
                                      ["vf", "v", "vvevu"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vv", "v", "vvvvu"],
+                                     ["vf", "v", "vvevu"]]>;
   }
 }
 
@@ -491,6 +499,10 @@ let HasMaskedOffOperand = false, Log2LMUL = [-2, -1, 0, 1, 2] in {
       defm ""  : RVVOutOp1Op2BuiltinSet<NAME, "x",
                                         [["vv", "w", "wwvv"],
                                          ["vf", "w", "wwev"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm ""  : RVVOutOp1Op2BuiltinSet<NAME, "y",
+                                        [["vv", "vw", "wwvv"],
+                                         ["vf", "vw", "wwev"]]>;
   }
   multiclass RVVFloatingWidenTerBuiltinSetRoundingMode {
     defm ""  : RVVOutOp1Op2BuiltinSet<NAME, "f",
@@ -500,10 +512,14 @@ let HasMaskedOffOperand = false, Log2LMUL = [-2, -1, 0, 1, 2] in {
       defm ""  : RVVOutOp1Op2BuiltinSet<NAME, "x",
                                         [["vv", "w", "wwvvu"],
                                          ["vf", "w", "wwevu"]]>;
+    let RequiredFeatures = ["zvfbfa"] in
+      defm ""  : RVVOutOp1Op2BuiltinSet<NAME, "y",
+                                        [["vv", "vw", "wwvvu"],
+                                         ["vf", "vw", "wwevu"]]>;
   }
 }
 
-multiclass RVVFloatingBinBuiltinSet {
+multiclass RVVFloatingBinBuiltinSet<bit HasBF = 0> {
   defm "" : RVVOutOp1BuiltinSet<NAME, "fd",
                                 [["vv", "v", "vvv"],
                                  ["vf", "v", "vve"]]>;
@@ -511,9 +527,15 @@ multiclass RVVFloatingBinBuiltinSet {
     defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                   [["vv", "v", "vvv"],
                                    ["vf", "v", "vve"]]>;
+  if HasBF then {
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vv", "v", "vvv"],
+                                     ["vf", "v", "vve"]]>;
+  }
 }
 
-multiclass RVVFloatingBinBuiltinSetRoundingMode {
+multiclass RVVFloatingBinBuiltinSetRoundingMode<bit HasBF = 0> {
   defm "" : RVVOutOp1BuiltinSet<NAME, "fd",
                                 [["vv", "v", "vvvu"],
                                  ["vf", "v", "vveu"]]>;
@@ -521,22 +543,38 @@ multiclass RVVFloatingBinBuiltinSetRoundingMode {
     defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                   [["vv", "v", "vvvu"],
                                    ["vf", "v", "vveu"]]>;
+  if HasBF then {
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vv", "v", "vvvu"],
+                                     ["vf", "v", "vveu"]]>;
+  }
 }
 
-multiclass RVVFloatingBinVFBuiltinSet {
+multiclass RVVFloatingBinVFBuiltinSet<bit HasBF = 0> {
   defm "" : RVVOutOp1BuiltinSet<NAME, "fd",
                                 [["vf", "v", "vve"]]>;
   let RequiredFeatures = ["zvfh"] in
     defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                   [["vf", "v", "vve"]]>;
+  if HasBF then {
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vf", "v", "vve"]]>;
+  }
 }
 
-multiclass RVVFloatingBinVFBuiltinSetRoundingMode {
+multiclass RVVFloatingBinVFBuiltinSetRoundingMode<bit HasBF = 0> {
   defm "" : RVVOutOp1BuiltinSet<NAME, "fd",
                                 [["vf", "v", "vveu"]]>;
   let RequiredFeatures = ["zvfh"] in
     defm "" : RVVOutOp1BuiltinSet<NAME, "x",
                                   [["vf", "v", "vveu"]]>;
+  if HasBF then {
+    let RequiredFeatures = ["zvfbfa"] in
+      defm "" : RVVOutOp1BuiltinSet<NAME, "y",
+                                    [["vf", "v", "vveu"]]>;
+  }
 }
 
 multiclass RVVFloatingMaskOutBuiltinSet {
@@ -547,6 +585,10 @@ multiclass RVVFloatingMaskOutBuiltinSet {
     defm "" : RVVOp0Op1BuiltinSet<NAME, "x",
                                   [["vv", "vm", "mvv"],
                                    ["vf", "vm", "mve"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm "" : RVVOp0Op1BuiltinSet<NAME, "y",
+                                  [["vv", "vm", "mvv"],
+                                   ["vf", "vm", "mve"]]>;
 }
 
 multiclass RVVFloatingMaskOutVFBuiltinSet
@@ -748,6 +790,10 @@ multiclass RVVFloatingWidenBinBuiltinSet {
     defm "" : RVVWidenBuiltinSet<NAME, "x",
                                  [["vv", "w", "wvv"],
                                   ["vf", "w", "wve"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm "" : RVVWidenBuiltinSet<NAME, "y",
+                                 [["vv", "vw", "wvv"],
+                                  ["vf", "vw", "wve"]]>;
 }
 
 multiclass RVVFloatingWidenBinBuiltinSetRoundingMode {
@@ -758,6 +804,10 @@ multiclass RVVFloatingWidenBinBuiltinSetRoundingMode {
     defm "" : RVVWidenBuiltinSet<NAME, "x",
                                  [["vv", "w", "wvvu"],
                                   ["vf", "w", "wveu"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm "" : RVVWidenBuiltinSet<NAME, "y",
+                                 [["vv", "vw", "wvvu"],
+                                  ["vf", "vw", "wveu"]]>;
 }
 
 multiclass RVVFloatingWidenOp0BinBuiltinSet {
@@ -768,6 +818,10 @@ multiclass RVVFloatingWidenOp0BinBuiltinSet {
     defm "" : RVVWidenWOp0BuiltinSet<NAME # "_w", "x",
                                      [["wv", "w", "wwv"],
                                       ["wf", "w", "wwe"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm "" : RVVWidenWOp0BuiltinSet<NAME # "_w", "y",
+                                     [["wv", "vw", "wwv"],
+                                      ["wf", "ew", "wwe"]]>;
 }
 
 multiclass RVVFloatingWidenOp0BinBuiltinSetRoundingMode {
@@ -778,4 +832,8 @@ multiclass RVVFloatingWidenOp0BinBuiltinSetRoundingMode {
     defm "" : RVVWidenWOp0BuiltinSet<NAME # "_w", "x",
                                      [["wv", "w", "wwvu"],
                                       ["wf", "w", "wweu"]]>;
+  let RequiredFeatures = ["zvfbfa"] in
+    defm "" : RVVWidenWOp0BuiltinSet<NAME # "_w", "y",
+                                     [["wv", "vw", "wwvu"],
+                                      ["wf", "ew", "wweu"]]>;
 }
diff --git a/clang/lib/Sema/SemaRISCV.cpp b/clang/lib/Sema/SemaRISCV.cpp
index 3ba93ff98898b..c5ef0d535628d 100644
--- a/clang/lib/Sema/SemaRISCV.cpp
+++ b/clang/lib/Sema/SemaRISCV.cpp
@@ -1464,7 +1464,8 @@ void SemaRISCV::checkRVVTypeSupport(QualType Ty, SourceLocation Loc, Decl *D,
     }
   else if (Info.ElementType->isBFloat16Type() &&
            !FeatureMap.lookup("zvfbfmin") &&
-           !FeatureMap.lookup("xandesvbfhcvt"))
+           !FeatureMap.lookup("xandesvbfhcvt") &&
+           !FeatureMap.lookup("experimental-zvfbfa"))
     if (DeclareAndesVectorBuiltins) {
       Diag(Loc, diag::err_riscv_type_requires_extension, D)
           << Ty << "zvfbfmin or xandesvbfhcvt";
diff --git a/clang/lib/Support/RISCVVIntrinsicUtils.cpp b/clang/lib/Support/RISCVVIntrinsicUtils.cpp
index 5a4e805d4a9d1..dad3d0dae423a 100644
--- a/clang/lib/Support/RISCVVIntrinsicUtils.cpp
+++ b/clang/lib/Support/RISCVVIntrinsicUtils.cpp
@@ -654,6 +654,9 @@ PrototypeDescriptor::parsePrototypeDescriptor(
     case 'F':
       TM |= TypeModifier::Float;
       break;
+    case 'Y':
+      TM |= TypeModifier::BFloat;
+      break;
     case 'S':
       TM |= TypeModifier::LMUL1;
       break;
@@ -704,6 +707,8 @@ void RVVType::applyModifier(const PrototypeDescriptor &Transformer) {
     ElementBitwidth *= 2;
     LMUL.MulLog2LMUL(1);
     Scale = LMUL.getScale(ElementBitwidth);
+    if (ScalarType == ScalarTypeKind::BFloat)
+      ScalarType = ScalarTypeKind::Float;
     break;
   case VectorTypeModifier::Widening4XVector:
     ElementBitwidth *= 4;
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfadd.c
new file mode 100644
index 0000000000000..d7734e05a8aaa
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfclass.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfclass.c
new file mode 100644
index 0000000000000..68814f4672d05
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfclass.c
@@ -0,0 +1,134 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1bf16.i64(<vscale x 1 x i16> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2bf16.i64(<vscale x 2 x i16> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4bf16.i64(<vscale x 4 x i16> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8bf16.i64(<vscale x 8 x i16> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16bf16.i64(<vscale x 16 x i16> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32bf16.i64(<vscale x 32 x i16> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8_m(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmacc.c
new file mode 100644
index 0000000000000..616455d5f3f9e
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmacc.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmadd.c
new file mode 100644
index 0000000000000..eec662a3671c8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmax.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmax.c
new file mode 100644
index 0000000000000..dfdeb4e967a46
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmax.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmerge.c
new file mode 100644
index 0000000000000..96221c5385dd9
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmerge.c
@@ -0,0 +1,69 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmerge_vfm_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmerge_vfm_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, vbool64_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16mf4(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmerge_vfm_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmerge_vfm_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, vbool32_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16mf2(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmerge_vfm_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmerge_vfm_bf16m1(vbfloat16m1_t op1, __bf16 op2, vbool16_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m1(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmerge_vfm_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmerge_vfm_bf16m2(vbfloat16m2_t op1, __bf16 op2, vbool8_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m2(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmerge_vfm_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmerge_vfm_bf16m4(vbfloat16m4_t op1, __bf16 op2, vbool4_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m4(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmerge_vfm_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmerge_vfm_bf16m8(vbfloat16m8_t op1, __bf16 op2, vbool2_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m8(op1, op2, mask, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmin.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmin.c
new file mode 100644
index 0000000000000..8f8d82ba21bc5
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmin.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsac.c
new file mode 100644
index 0000000000000..f4644dfb8d7e7
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsac.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsub.c
new file mode 100644
index 0000000000000..07053afa3355c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmul.c
new file mode 100644
index 0000000000000..88fb329934365
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmul.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmv.c
new file mode 100644
index 0000000000000..d80ec3df1bbaa
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfmv.c
@@ -0,0 +1,189 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_v_f_bf16mf4(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16.i64(<vscale x 1 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_v_f_bf16mf4(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16mf4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_v_f_bf16mf2(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16.i64(<vscale x 2 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_v_f_bf16mf2(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16mf2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_v_f_bf16m1(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16.i64(<vscale x 4 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_v_f_bf16m1(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m1(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_v_f_bf16m2(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16.i64(<vscale x 8 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_v_f_bf16m2(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_v_f_bf16m4(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16.i64(<vscale x 16 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_v_f_bf16m4(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_v_f_bf16m8(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16.i64(<vscale x 32 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_v_f_bf16m8(__bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m8(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16mf4_bf16(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv1bf16(<vscale x 1 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16mf4_bf16(vbfloat16mf4_t src) {
+  return __riscv_vfmv_f_s_bf16mf4_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_s_f_bf16mf4(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16.i64(<vscale x 1 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_s_f_bf16mf4(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16mf4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16mf2_bf16(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv2bf16(<vscale x 2 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16mf2_bf16(vbfloat16mf2_t src) {
+  return __riscv_vfmv_f_s_bf16mf2_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_s_f_bf16mf2(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16.i64(<vscale x 2 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_s_f_bf16mf2(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16mf2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m1_bf16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv4bf16(<vscale x 4 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m1_bf16(vbfloat16m1_t src) {
+  return __riscv_vfmv_f_s_bf16m1_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_s_f_bf16m1(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16.i64(<vscale x 4 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_s_f_bf16m1(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m1(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m2_bf16(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv8bf16(<vscale x 8 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m2_bf16(vbfloat16m2_t src) {
+  return __riscv_vfmv_f_s_bf16m2_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_s_f_bf16m2(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16.i64(<vscale x 8 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_s_f_bf16m2(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m2(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m4_bf16(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv16bf16(<vscale x 16 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m4_bf16(vbfloat16m4_t src) {
+  return __riscv_vfmv_f_s_bf16m4_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_s_f_bf16m4(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16.i64(<vscale x 16 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_s_f_bf16m4(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m4(src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m8_bf16(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv32bf16(<vscale x 32 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m8_bf16(vbfloat16m8_t src) {
+  return __riscv_vfmv_f_s_bf16m8_bf16(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_s_f_bf16m8(
+// CHECK-RV64-SAME: bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16.i64(<vscale x 32 x bfloat> poison, bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_s_f_bf16m8(__bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m8(src, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt.c
new file mode 100644
index 0000000000000..a5afab9bec1ec
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt.c
@@ -0,0 +1,724 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_m(vbool32_t vm, vfloat32m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_m(vbool64_t vm,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_m(vbool32_t vm,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_m(vbool64_t vm,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_m(vbool32_t vm,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_m(vbool16_t vm,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm_m(vm, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rod.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rod.c
new file mode 100644
index 0000000000000..70c377bba1dfb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rod.c
@@ -0,0 +1,113 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_m(vbool16_t vm, vfloat32m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_m(vbool8_t vm, vfloat32m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_m(vbool4_t vm, vfloat32m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4_m(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rtz.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rtz.c
new file mode 100644
index 0000000000000..854e9868109e4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfncvt_rtz.c
@@ -0,0 +1,267 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8(vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4(vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_m(vbool32_t vm,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_m(vbool64_t vm,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_m(vbool32_t vm,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_m(vbool16_t vm,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4_m(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmacc.c
new file mode 100644
index 0000000000000..18484883a14f4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmacc.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmadd.c
new file mode 100644
index 0000000000000..e519e5acb4575
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsac.c
new file mode 100644
index 0000000000000..47e1f44f5a45f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsac.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsub.c
new file mode 100644
index 0000000000000..4b55b64542c61
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfnmsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrec7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrec7.c
new file mode 100644
index 0000000000000..1ffee73e91d04
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrec7.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4(vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2(vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1(vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2(vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4(vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8(vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8_m(mask, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsqrt7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsqrt7.c
new file mode 100644
index 0000000000000..964c4869622aa
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsqrt7.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4(vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2(vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1(vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2(vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4(vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8(vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4_m(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8_m(mask, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsub.c
new file mode 100644
index 0000000000000..c7c3869e7b77c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfrsub.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnj.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnj.c
new file mode 100644
index 0000000000000..778b8b83e9841
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnj.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjn.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjn.c
new file mode 100644
index 0000000000000..7de308978e1d3
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjn.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjx.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjx.c
new file mode 100644
index 0000000000000..5fa285cc78b63
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsgnjx.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1down.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1down.c
new file mode 100644
index 0000000000000..b94d26b4ddf40
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1down.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4(vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2(vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1(vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2(vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4(vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8(vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8_m(mask, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1up.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1up.c
new file mode 100644
index 0000000000000..06e8b49af19d0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfslide1up.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4(vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2(vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1(vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2(vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4(vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8(vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4_m(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8_m(mask, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsub.c
new file mode 100644
index 0000000000000..2423b0bbdbb80
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwadd.c
new file mode 100644
index 0000000000000..24d34f46f4203
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwadd.c
@@ -0,0 +1,899 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2(vfloat32mf2_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2(vfloat32mf2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1(vfloat32m1_t vs2, vbfloat16mf2_t vs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1(vfloat32m1_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2(vfloat32m2_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4(vfloat32m4_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8(vfloat32m8_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm(vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm(vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm(vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm(vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm(vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm(vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm(vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwcvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwcvt.c
new file mode 100644
index 0000000000000..fb3e0031af98c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwcvt.c
@@ -0,0 +1,366 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4(vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2(vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1(vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2(vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4(vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8(vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4(vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2(vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1(vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2(vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4(vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8(vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_m(vbool64_t vm, vint8mf8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_m(vbool32_t vm, vint8mf4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_m(vbool16_t vm, vint8mf2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_m(vbool8_t vm, vint8m1_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_m(vbool4_t vm, vint8m2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_m(vbool2_t vm, vint8m4_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_m(vbool64_t vm, vuint8mf8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_m(vbool32_t vm, vuint8mf4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_m(vbool16_t vm, vuint8mf2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_m(vbool8_t vm, vuint8m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_m(vbool4_t vm, vuint8m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_m(vbool2_t vm, vuint8m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4_m(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8_m(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmacc.c
new file mode 100644
index 0000000000000..be09003320386
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmacc.c
@@ -0,0 +1,486 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                             vbfloat16mf4_t vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               __bf16 vs1, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             __bf16 vs1, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmsac.c
new file mode 100644
index 0000000000000..749081333c2b3
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmsac.c
@@ -0,0 +1,486 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                             vbfloat16mf4_t vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               __bf16 vs1, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             __bf16 vs1, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmul.c
new file mode 100644
index 0000000000000..6783ba43b0570
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwmul.c
@@ -0,0 +1,455 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmacc.c
new file mode 100644
index 0000000000000..6127a94c919d9
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmacc.c
@@ -0,0 +1,494 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmsac.c
new file mode 100644
index 0000000000000..f37dd310d944d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwnmsac.c
@@ -0,0 +1,494 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_m(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm_m(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm_m(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm_m(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwsub.c
new file mode 100644
index 0000000000000..510ff9193389e
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vfwsub.c
@@ -0,0 +1,899 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2(vfloat32mf2_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2(vfloat32mf2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1(vfloat32m1_t vs2, vbfloat16mf2_t vs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1(vfloat32m1_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2(vfloat32m2_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4(vfloat32m4_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8(vfloat32m8_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_m(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_m(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm(vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm(vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm(vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm(vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm(vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm(vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm(vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm_m(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm_m(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfeq.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfeq.c
new file mode 100644
index 0000000000000..669d0427b569a
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfeq.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfge.c
new file mode 100644
index 0000000000000..b169efd51a0b4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfge.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfgt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfgt.c
new file mode 100644
index 0000000000000..9aea7d24b0edc
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfgt.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfle.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfle.c
new file mode 100644
index 0000000000000..40f0c27f5b37a
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfle.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmflt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmflt.c
new file mode 100644
index 0000000000000..f64eee3effdaf
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmflt.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfne.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfne.c
new file mode 100644
index 0000000000000..809ea5628e394
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/non-overloaded/vmfne.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf4_b64(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf2_b32(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m1_b16(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m2_b8(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m4_b4(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m8_b2(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf4_b64_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf2_b32_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m1_b16_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m2_b8_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m4_b4_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m8_b2_m(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfadd.c
new file mode 100644
index 0000000000000..9d6b071c2768c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfclass.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfclass.c
new file mode 100644
index 0000000000000..2760f85a45d3c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfclass.c
@@ -0,0 +1,134 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1bf16.i64(<vscale x 1 x i16> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2bf16.i64(<vscale x 2 x i16> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4bf16.i64(<vscale x 4 x i16> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8bf16.i64(<vscale x 8 x i16> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16bf16.i64(<vscale x 16 x i16> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32bf16.i64(<vscale x 32 x i16> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfclass(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmacc.c
new file mode 100644
index 0000000000000..ae3f1f24eb762
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmacc.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmadd.c
new file mode 100644
index 0000000000000..db2184c6e1dba
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmax.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmax.c
new file mode 100644
index 0000000000000..66497bfb15934
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmax.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmerge.c
new file mode 100644
index 0000000000000..1dc290bf1a222
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmerge.c
@@ -0,0 +1,69 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmerge_vfm_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmerge_vfm_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, vbool64_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmerge_vfm_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmerge_vfm_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, vbool32_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmerge_vfm_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmerge_vfm_bf16m1(vbfloat16m1_t op1, __bf16 op2, vbool16_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmerge_vfm_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmerge_vfm_bf16m2(vbfloat16m2_t op1, __bf16 op2, vbool8_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmerge_vfm_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmerge_vfm_bf16m4(vbfloat16m4_t op1, __bf16 op2, vbool4_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmerge_vfm_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmerge_vfm_bf16m8(vbfloat16m8_t op1, __bf16 op2, vbool2_t mask, size_t vl) {
+  return __riscv_vfmerge(op1, op2, mask, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmin.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmin.c
new file mode 100644
index 0000000000000..1564d1195ecef
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmin.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsac.c
new file mode 100644
index 0000000000000..0384e7da29e08
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsac.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsub.c
new file mode 100644
index 0000000000000..306f189b54c89
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmul.c
new file mode 100644
index 0000000000000..fffd83a12d36c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmul.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmv.c
new file mode 100644
index 0000000000000..f85378ff0f37a
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfmv.c
@@ -0,0 +1,69 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16mf4_bf16(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv1bf16(<vscale x 1 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16mf4_bf16(vbfloat16mf4_t src) {
+  return __riscv_vfmv_f(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16mf2_bf16(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv2bf16(<vscale x 2 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16mf2_bf16(vbfloat16mf2_t src) {
+  return __riscv_vfmv_f(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m1_bf16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv4bf16(<vscale x 4 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m1_bf16(vbfloat16m1_t src) {
+  return __riscv_vfmv_f(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m2_bf16(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv8bf16(<vscale x 8 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m2_bf16(vbfloat16m2_t src) {
+  return __riscv_vfmv_f(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m4_bf16(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv16bf16(<vscale x 16 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m4_bf16(vbfloat16m4_t src) {
+  return __riscv_vfmv_f(src);
+}
+
+// CHECK-RV64-LABEL: define dso_local bfloat @test_vfmv_f_s_bf16m8_bf16(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call bfloat @llvm.riscv.vfmv.f.s.nxv32bf16(<vscale x 32 x bfloat> [[SRC]])
+// CHECK-RV64-NEXT:    ret bfloat [[TMP0]]
+//
+__bf16 test_vfmv_f_s_bf16m8_bf16(vbfloat16m8_t src) {
+  return __riscv_vfmv_f(src);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt.c
new file mode 100644
index 0000000000000..fb635d61670c2
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt.c
@@ -0,0 +1,724 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_m(vbool32_t vm, vfloat32m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16(vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_m(vbool64_t vm,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_m(vbool32_t vm,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_m(vbool64_t vm,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_m(vbool32_t vm,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_m(vbool16_t vm,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_xu(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_bf16(vm, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rod.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rod.c
new file mode 100644
index 0000000000000..1ad856df48c4b
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rod.c
@@ -0,0 +1,113 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4(vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2(vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1(vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2(vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4(vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_m(vbool16_t vm, vfloat32m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_m(vbool8_t vm, vfloat32m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_m(vbool4_t vm, vfloat32m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rtz.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rtz.c
new file mode 100644
index 0000000000000..12d08934bd49d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfncvt_rtz.c
@@ -0,0 +1,267 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8(vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4(vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4(vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_m(vbool32_t vm,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_rtz_x(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_m(vbool64_t vm,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_m(vbool32_t vm,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_m(vbool16_t vm,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> poison, <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_m(vbool2_t vm, vbfloat16m8_t vs2,
+                                                size_t vl) {
+  return __riscv_vfncvt_rtz_xu(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmacc.c
new file mode 100644
index 0000000000000..6f7928b763230
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmacc.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmadd.c
new file mode 100644
index 0000000000000..97d207040e5b2
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmadd.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsac.c
new file mode 100644
index 0000000000000..404b4f83c4f1f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsac.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsub.c
new file mode 100644
index 0000000000000..3a520dd9af9bf
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfnmsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrec7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrec7.c
new file mode 100644
index 0000000000000..462b6acf8eee6
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrec7.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4(vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2(vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1(vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2(vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4(vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8(vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7(mask, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsqrt7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsqrt7.c
new file mode 100644
index 0000000000000..051fde7ef5472
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsqrt7.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4(vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2(vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1(vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2(vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4(vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8(vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7(mask, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsub.c
new file mode 100644
index 0000000000000..04941823d1917
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfrsub.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnj.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnj.c
new file mode 100644
index 0000000000000..615deddf4254f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnj.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjn.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjn.c
new file mode 100644
index 0000000000000..a895e5f6ad6bd
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjn.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjx.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjx.c
new file mode 100644
index 0000000000000..0187516d00dc1
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsgnjx.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1down.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1down.c
new file mode 100644
index 0000000000000..4a7689492de92
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1down.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4(vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2(vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1(vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2(vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4(vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8(vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down(mask, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1up.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1up.c
new file mode 100644
index 0000000000000..f9f2dc0cf26b8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfslide1up.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4(vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2(vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1(vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2(vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4(vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8(vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up(mask, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsub.c
new file mode 100644
index 0000000000000..ebcf6fa4669a3
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwadd.c
new file mode 100644
index 0000000000000..124e7fb6de342
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwadd.c
@@ -0,0 +1,893 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2(vfloat32mf2_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2(vfloat32mf2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1(vfloat32m1_t vs2, vbfloat16mf2_t vs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1(vfloat32m1_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2(vfloat32m2_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4(vfloat32m4_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8(vfloat32m8_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm(vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm(vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm(vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm(vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm(vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm(vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm(vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwcvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwcvt.c
new file mode 100644
index 0000000000000..0399a639898a0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwcvt.c
@@ -0,0 +1,366 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4(vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2(vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1(vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2(vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4(vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8(vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4(vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2(vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1(vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2(vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4(vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8(vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2(vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1(vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2(vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4(vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8(vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f(vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_m(vbool64_t vm, vint8mf8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_m(vbool32_t vm, vint8mf4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_m(vbool16_t vm, vint8mf2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_m(vbool8_t vm, vint8m1_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_m(vbool4_t vm, vint8m2_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_m(vbool2_t vm, vint8m4_t vs2,
+                                         size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> poison, <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_m(vbool64_t vm, vuint8mf8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> poison, <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_m(vbool32_t vm, vuint8mf4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> poison, <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_m(vbool16_t vm, vuint8mf2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> poison, <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_m(vbool8_t vm, vuint8m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> poison, <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_m(vbool4_t vm, vuint8m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> poison, <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_m(vbool2_t vm, vuint8m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwcvt_f(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwcvt_f(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f(vm, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwcvt_f(vm, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmacc.c
new file mode 100644
index 0000000000000..2eb7fc8cdeea8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmacc.c
@@ -0,0 +1,474 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                             vbfloat16mf4_t vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               __bf16 vs1, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             __bf16 vs1, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmsac.c
new file mode 100644
index 0000000000000..28f507619b428
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmsac.c
@@ -0,0 +1,474 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                             vbfloat16mf4_t vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                             vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                           vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                          vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                          vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                          vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs1,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                               __bf16 vs1, vbfloat16mf4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs1,
+                                             vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                             __bf16 vs1, vbfloat16mf2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs1,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                            __bf16 vs1, vbfloat16m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs1,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                            __bf16 vs1, vbfloat16m2_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs1,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                            __bf16 vs1, vbfloat16m4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmul.c
new file mode 100644
index 0000000000000..8de49fa586dcf
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwmul.c
@@ -0,0 +1,451 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmacc.c
new file mode 100644
index 0000000000000..783693106a217
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmacc.c
@@ -0,0 +1,480 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmsac.c
new file mode 100644
index 0000000000000..ca936af7140e5
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwnmsac.c
@@ -0,0 +1,480 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2(vfloat32mf2_t vd,
+                                              vbfloat16mf4_t vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2(vfloat32mf2_t vd, __bf16 vs1,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1(vfloat32m1_t vd, vbfloat16mf2_t vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1(vfloat32m1_t vd, __bf16 vs1,
+                                            vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2(vfloat32m2_t vd, __bf16 vs1,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4(vfloat32m4_t vd, __bf16 vs1,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8(vfloat32m8_t vd, __bf16 vs1,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwsub.c
new file mode 100644
index 0000000000000..2e22e22e4b6fc
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vfwsub.c
@@ -0,0 +1,893 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2(vbfloat16mf4_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2(vbfloat16mf4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2(vfloat32mf2_t vs2,
+                                            vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2(vfloat32mf2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1(vbfloat16mf2_t vs2,
+                                          vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1(vbfloat16mf2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1(vfloat32m1_t vs2, vbfloat16mf2_t vs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1(vfloat32m1_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2(vbfloat16m1_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2(vbfloat16m1_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2(vfloat32m2_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4(vbfloat16m2_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4(vbfloat16m2_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4(vfloat32m4_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8(vbfloat16m4_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8(vbfloat16m4_t vs2, __bf16 rs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                         size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8(vfloat32m8_t vs2, __bf16 rs1,
+                                       size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_m(vbool64_t vm, vbfloat16mf4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_m(vbool32_t vm, vfloat32m1_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                           vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_m(vbool16_t vm, vfloat32m2_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                           vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_m(vbool8_t vm, vfloat32m4_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                           __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                           vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_m(vbool4_t vm, vfloat32m8_t vs2,
+                                         __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm(vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm(vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm(vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm(vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm(vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm(vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm(vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm(vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm(vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm(vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm(vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm(vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm(vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm(vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm(vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm(vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm(vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv(vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm(vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf(vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_m(vbool64_t vm,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> poison, <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_m(vbool64_t vm, vfloat32mf2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_m(vbool32_t vm, vbfloat16mf2_t vs2,
+                                               __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> poison, <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_m(vbool32_t vm, vfloat32m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_m(vbool16_t vm, vbfloat16m1_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> poison, <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_m(vbool16_t vm, vfloat32m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_m(vbool8_t vm, vbfloat16m2_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> poison, <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_m(vbool8_t vm, vfloat32m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_m(vbool4_t vm, vbfloat16m4_t vs2,
+                                              __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv(vm, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> poison, <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 3)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_m(vbool4_t vm, vfloat32m8_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf(vm, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfeq.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfeq.c
new file mode 100644
index 0000000000000..29881c9ba5d5e
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfeq.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfge.c
new file mode 100644
index 0000000000000..b8083c5ebb8df
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfge.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfgt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfgt.c
new file mode 100644
index 0000000000000..b8749b3295a19
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfgt.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfle.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfle.c
new file mode 100644
index 0000000000000..724608c89e8de
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfle.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmflt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmflt.c
new file mode 100644
index 0000000000000..1b0b898d5053f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmflt.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfne.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfne.c
new file mode 100644
index 0000000000000..672c1504d73eb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/non-policy/overloaded/vmfne.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64(vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64(vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32(vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32(vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16(vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16(vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8(vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8(vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4(vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4(vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2(vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2(vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64_m(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> poison, <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64_m(vbool64_t mask, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32_m(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> poison, <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32_m(vbool32_t mask, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16_m(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> poison, <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16_m(vbool16_t mask, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8_m(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> poison, <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8_m(vbool8_t mask, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4_m(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> poison, <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4_m(vbool4_t mask, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2_m(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> poison, <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2_m(vbool2_t mask, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne(mask, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfadd.c
new file mode 100644
index 0000000000000..6d55279ae306f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfclass.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfclass.c
new file mode 100644
index 0000000000000..8e6946de398b0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfclass.c
@@ -0,0 +1,272 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tu(vuint16mf4_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tu(vuint16mf2_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tu(vuint16m1_t vd, vbfloat16m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tu(vuint16m2_t vd, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tu(vuint16m4_t vd, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tu(vuint16m8_t vd, vbfloat16m8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tum(vbool64_t vm, vuint16mf4_t vd,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tum(vbool32_t vm, vuint16mf2_t vd,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tum(vbool16_t vm, vuint16m1_t vd,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tum(vbool8_t vm, vuint16m2_t vd,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tum(vbool4_t vm, vuint16m4_t vd,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tum(vbool2_t vm, vuint16m8_t vd,
+                                            vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tumu(vbool64_t vm, vuint16mf4_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tumu(vbool32_t vm, vuint16mf2_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tumu(vbool16_t vm, vuint16m1_t vd,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tumu(vbool8_t vm, vuint16m2_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tumu(vbool4_t vm, vuint16m4_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tumu(vbool2_t vm, vuint16m8_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_mu(vbool64_t vm, vuint16mf4_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf4_u16mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_mu(vbool32_t vm, vuint16mf2_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16mf2_u16mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_mu(vbool16_t vm, vuint16m1_t vd,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m1_u16m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_mu(vbool8_t vm, vuint16m2_t vd,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m2_u16m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_mu(vbool4_t vm, vuint16m4_t vd,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m4_u16m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_mu(vbool2_t vm, vuint16m8_t vd,
+                                           vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_v_bf16m8_u16m8_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmacc.c
new file mode 100644
index 0000000000000..2d4e481b1f8bb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmacc.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmadd.c
new file mode 100644
index 0000000000000..511e073e10143
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmax.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmax.c
new file mode 100644
index 0000000000000..f3698d41aff34
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmax.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmerge.c
new file mode 100644
index 0000000000000..bcaf2cb8817e4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmerge.c
@@ -0,0 +1,69 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmerge_vfm_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmerge_vfm_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, vbool64_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16mf4_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmerge_vfm_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmerge_vfm_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, vbool32_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16mf2_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmerge_vfm_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmerge_vfm_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, vbool16_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m1_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmerge_vfm_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmerge_vfm_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, vbool8_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m2_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmerge_vfm_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmerge_vfm_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, vbool4_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m4_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmerge_vfm_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmerge_vfm_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, vbool2_t mask, size_t vl) {
+  return __riscv_vfmerge_vfm_bf16m8_tu(maskedoff, op1, op2, mask, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmin.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmin.c
new file mode 100644
index 0000000000000..911f8792e4436
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmin.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsac.c
new file mode 100644
index 0000000000000..9575ad337de8c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsac.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsub.c
new file mode 100644
index 0000000000000..8e382f710ab5a
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmul.c
new file mode 100644
index 0000000000000..716f056f3e12c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmul.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmv.c
new file mode 100644
index 0000000000000..069ee6a2a5df3
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfmv.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_v_f_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_v_f_bf16mf4_tu(vbfloat16mf4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16mf4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_v_f_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_v_f_bf16mf2_tu(vbfloat16mf2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16mf2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_v_f_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_v_f_bf16m1_tu(vbfloat16m1_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m1_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_v_f_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_v_f_bf16m2_tu(vbfloat16m2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_v_f_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_v_f_bf16m4_tu(vbfloat16m4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_v_f_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_v_f_bf16m8_tu(vbfloat16m8_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_f_bf16m8_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_s_f_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_s_f_bf16mf4_tu(vbfloat16mf4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16mf4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_s_f_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_s_f_bf16mf2_tu(vbfloat16mf2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16mf2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_s_f_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_s_f_bf16m1_tu(vbfloat16m1_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m1_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_s_f_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_s_f_bf16m2_tu(vbfloat16m2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m2_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_s_f_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_s_f_bf16m4_tu(vbfloat16m4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m4_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_s_f_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_s_f_bf16m8_tu(vbfloat16m8_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_f_bf16m8_tu(maskedoff, src, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt.c
new file mode 100644
index 0000000000000..36d4fc332499f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt.c
@@ -0,0 +1,1577 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tu(vint8mf8_t vd, vbfloat16mf4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tu(vint8mf4_t vd, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tu(vint8mf2_t vd, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tu(vuint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tu(vuint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tu(vuint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tu(vuint8m1_t vd, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tu(vuint8m2_t vd, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tu(vuint8m4_t vd, vbfloat16m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tu(vbfloat16mf4_t vd,
+                                            vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tu(vbfloat16mf2_t vd, vfloat32m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tu(vbfloat16m1_t vd, vfloat32m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tu(vbfloat16m2_t vd, vfloat32m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tu(vbfloat16m4_t vd, vfloat32m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tum(vbool64_t vm, vint8mf8_t vd,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tum(vbool32_t vm, vint8mf4_t vd,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tum(vbool16_t vm, vint8mf2_t vd,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tum(vbool8_t vm, vint8m1_t vd,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tum(vbool4_t vm, vint8m2_t vd,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tum(vbool2_t vm, vint8m4_t vd,
+                                            vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tum(vbool64_t vm, vuint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tum(vbool32_t vm, vuint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tum(vbool16_t vm, vuint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tum(vbool8_t vm, vuint8m1_t vd,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tum(vbool4_t vm, vuint8m2_t vd,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tum(vbool2_t vm, vuint8m4_t vd,
+                                              vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                           vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                           vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                           vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tumu(vbool16_t vm, vint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tumu(vbool8_t vm, vint8m1_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tumu(vbool4_t vm, vint8m2_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tumu(vbool2_t vm, vint8m4_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tumu(vbool64_t vm, vuint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tumu(vbool32_t vm, vuint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tumu(vbool16_t vm, vuint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tumu(vbool8_t vm, vuint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tumu(vbool4_t vm, vuint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tumu(vbool2_t vm, vuint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                            vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                            vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                            vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_mu(vbool64_t vm, vint8mf8_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_mu(vbool32_t vm, vint8mf4_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_mu(vbool16_t vm, vint8mf2_t vd,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_mu(vbool8_t vm, vint8m1_t vd,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_mu(vbool4_t vm, vint8m2_t vd,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_mu(vbool2_t vm, vint8m4_t vd,
+                                           vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_mu(vbool64_t vm, vuint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_mu(vbool32_t vm, vuint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_mu(vbool16_t vm, vuint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_mu(vbool8_t vm, vuint8m1_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_mu(vbool4_t vm, vuint8m2_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_mu(vbool2_t vm, vuint8m4_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                            vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                            vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                          vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                          vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                          vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tu(vint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tu(vint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tu(vint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tu(vuint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tu(vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tu(vuint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tu(vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tu(vuint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tu(vuint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tu(vuint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tu(vuint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tu(vbfloat16mf4_t vd,
+                                               vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tu(vbfloat16mf2_t vd,
+                                               vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tu(vbfloat16m1_t vd, vfloat32m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tu(vbfloat16m2_t vd, vfloat32m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tu(vbfloat16m4_t vd, vfloat32m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tum(vbool64_t vm, vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tum(vbool32_t vm, vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tum(vbool16_t vm, vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tum(vbool8_t vm, vint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tum(vbool4_t vm, vint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tum(vbool2_t vm, vint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tum(vbool64_t vm,
+                                                    vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tum(vm, vd, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tum(vbool32_t vm,
+                                                    vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tum(vm, vd, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tum(vbool16_t vm, vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tum(vbool8_t vm, vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tum(vbool4_t vm, vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tum(vbool2_t vm, vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tumu(vm, vd, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tumu(vm, vd, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tumu(vbool16_t vm, vint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tumu(vbool8_t vm, vint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tumu(vbool4_t vm, vint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tumu(vbool2_t vm, vint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tumu(vbool64_t vm,
+                                                     vuint8mf8_t vd,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tumu(vm, vd, vs2,
+                                                     __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tumu(vbool32_t vm,
+                                                     vuint8mf4_t vd,
+                                                     vbfloat16mf2_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tumu(vm, vd, vs2,
+                                                     __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tumu(vbool16_t vm,
+                                                    vuint8mf2_t vd,
+                                                    vbfloat16m1_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tumu(vm, vd, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tumu(vbool8_t vm, vuint8m1_t vd,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tumu(vbool4_t vm, vuint8m2_t vd,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tumu(vbool2_t vm, vuint8m4_t vd,
+                                                  vbfloat16m8_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tumu(vbool64_t vm,
+                                                 vbfloat16mf4_t vd,
+                                                 vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tumu(vbool32_t vm,
+                                                 vbfloat16mf2_t vd,
+                                                 vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                               vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                               vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                               vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_mu(vbool64_t vm, vint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf4_i8mf8_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_mu(vbool32_t vm, vint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16mf2_i8mf4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_mu(vbool16_t vm, vint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m1_i8mf2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_mu(vbool8_t vm, vint8m1_t vd,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m2_i8m1_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_mu(vbool4_t vm, vint8m2_t vd,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m4_i8m2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_mu(vbool2_t vm, vint8m4_t vd,
+                                              vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_f_w_bf16m8_i8m4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_mu(vbool64_t vm, vuint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_mu(vbool32_t vm, vuint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                   vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_mu(vbool16_t vm, vuint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m1_u8mf2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_mu(vbool8_t vm, vuint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m2_u8m1_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_mu(vbool4_t vm, vuint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m4_u8m2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_mu(vbool2_t vm, vuint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_f_w_bf16m8_u8m4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                               vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                               vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16mf2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                             vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m1_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                             vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m2_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                             vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_f_w_bf16m4_rm_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rod.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rod.c
new file mode 100644
index 0000000000000..84066846e188c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rod.c
@@ -0,0 +1,233 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tu(vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tu(vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tu(vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tu(vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tu(vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tum(vbool64_t vm,
+                                                 vbfloat16mf4_t vd,
+                                                 vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tum(vbool32_t vm,
+                                                 vbfloat16mf2_t vd,
+                                                 vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                               vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                               vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                               vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tumu(vbool64_t vm,
+                                                  vbfloat16mf4_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tumu(vbool32_t vm,
+                                                  vbfloat16mf2_t vd,
+                                                  vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                                vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                                vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                                vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_f_w_bf16m4_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rtz.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rtz.c
new file mode 100644
index 0000000000000..4644eff12e210
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfncvt_rtz.c
@@ -0,0 +1,572 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tu(vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tu(vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tu(vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tu(vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tu(vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tu(vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tu(vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tu(vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tu(vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tum(vbool64_t vm, vint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tum(vbool32_t vm, vint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tum(vbool16_t vm, vint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tum(vbool8_t vm, vint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tum(vbool4_t vm, vint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tum(vbool2_t vm, vint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tum(vbool64_t vm,
+                                                     vuint8mf8_t vd,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tum(vbool32_t vm,
+                                                     vuint8mf4_t vd,
+                                                     vbfloat16mf2_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tum(vbool16_t vm,
+                                                    vuint8mf2_t vd,
+                                                    vbfloat16m1_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tum(vbool8_t vm, vuint8m1_t vd,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tum(vbool4_t vm, vuint8m2_t vd,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tum(vbool2_t vm, vuint8m4_t vd,
+                                                  vbfloat16m8_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tumu(vbool16_t vm, vint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tumu(vbool8_t vm, vint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tumu(vbool4_t vm, vint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tumu(vbool2_t vm, vint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tumu(vbool64_t vm,
+                                                      vuint8mf8_t vd,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tumu(vbool32_t vm,
+                                                      vuint8mf4_t vd,
+                                                      vbfloat16mf2_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tumu(vbool16_t vm,
+                                                     vuint8mf2_t vd,
+                                                     vbfloat16m1_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tumu(vbool8_t vm, vuint8m1_t vd,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tumu(vbool4_t vm, vuint8m2_t vd,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tumu(vbool2_t vm, vuint8m4_t vd,
+                                                   vbfloat16m8_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_mu(vbool64_t vm, vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_mu(vbool32_t vm, vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_mu(vbool16_t vm, vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m1_i8mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_mu(vbool8_t vm, vint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m2_i8m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_mu(vbool4_t vm, vint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m4_i8m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_mu(vbool2_t vm, vint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_f_w_bf16m8_i8m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_mu(vbool64_t vm,
+                                                    vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_mu(vbool32_t vm,
+                                                    vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_mu(vbool16_t vm, vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_mu(vbool8_t vm, vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m2_u8m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_mu(vbool4_t vm, vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m4_u8m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_mu(vbool2_t vm, vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_f_w_bf16m8_u8m4_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmacc.c
new file mode 100644
index 0000000000000..93fd6ba2faff7
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmacc.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmadd.c
new file mode 100644
index 0000000000000..d7e6b8225d8c0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsac.c
new file mode 100644
index 0000000000000..e0c289d23c17f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsac.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsub.c
new file mode 100644
index 0000000000000..05ccda36e9032
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfnmsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16mf2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16mf2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m1_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m1_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m2_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m2_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m4_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m4_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vv_bf16m8_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_vf_bf16m8_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrec7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrec7.c
new file mode 100644
index 0000000000000..3123692b45a68
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrec7.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf4_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16mf2_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m1_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m2_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m4_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_v_bf16m8_mu(mask, maskedoff, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsqrt7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsqrt7.c
new file mode 100644
index 0000000000000..8436f0ea488b2
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsqrt7.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf4_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16mf2_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m1_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m2_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m4_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_v_bf16m8_mu(mask, maskedoff, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsub.c
new file mode 100644
index 0000000000000..7dd2bb6cc502d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfrsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnj.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnj.c
new file mode 100644
index 0000000000000..b39a0be2238a9
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnj.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjn.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjn.c
new file mode 100644
index 0000000000000..7542e7866c888
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjn.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjx.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjx.c
new file mode 100644
index 0000000000000..104149e1b2fa4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsgnjx.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1down.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1down.c
new file mode 100644
index 0000000000000..228dc1cd064a8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1down.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf4_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16mf2_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m1_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m2_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m4_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_vf_bf16m8_mu(mask, maskedoff, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1up.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1up.c
new file mode 100644
index 0000000000000..9e6ff2b16f77d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfslide1up.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf4_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16mf2_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m1_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m2_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m4_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_vf_bf16m8_mu(mask, maskedoff, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsub.c
new file mode 100644
index 0000000000000..b6fd94ece20cd
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16mf2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m1_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_vv_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_vf_bf16m8_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwadd.c
new file mode 100644
index 0000000000000..4bee376cfe0fb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwadd.c
@@ -0,0 +1,2007 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tu(vfloat32mf2_t vd, vfloat32mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             vfloat32mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           vfloat32m1_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           vfloat32m2_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           vfloat32m4_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           vfloat32m8_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              vfloat32mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            vfloat32m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vfloat32mf2_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vfloat32m1_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm_tum(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vfloat32m2_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vfloat32m4_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vfloat32m8_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, vs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, rs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vfloat32mf2_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, vs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm_tumu(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vfloat32m1_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vfloat32m2_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vfloat32m4_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vfloat32m8_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32mf2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16mf2_f32m1_rm_mu(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m1_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m1_f32m2_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m1_f32m2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m1_f32m2_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m2_f32m4_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m2_f32m4_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m2_f32m4_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m4_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_bf16m4_f32m8_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_bf16m4_f32m8_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_bf16m4_f32m8_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_bf16_f32m8_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwcvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwcvt.c
new file mode 100644
index 0000000000000..9151319fcfb17
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwcvt.c
@@ -0,0 +1,765 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tu(vbfloat16mf4_t vd, vint8mf8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tu(vbfloat16mf2_t vd, vint8mf4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tu(vbfloat16m1_t vd, vint8mf2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tu(vbfloat16m2_t vd, vint8m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tu(vbfloat16m4_t vd, vint8m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tu(vbfloat16m8_t vd, vint8m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tu(vbfloat16mf4_t vd, vuint8mf8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tu(vbfloat16mf2_t vd, vuint8mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tu(vbfloat16m1_t vd, vuint8mf2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tu(vbfloat16m2_t vd, vuint8m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tu(vbfloat16m4_t vd, vuint8m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tu(vbfloat16m8_t vd, vuint8m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                           vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                           vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                           vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tum(vbool2_t vm, vbfloat16m8_t vd,
+                                           vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                            vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                            vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                            vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tum(vbool2_t vm, vbfloat16m8_t vd,
+                                            vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                            vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                            vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                            vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tumu(vbool2_t vm, vbfloat16m8_t vd,
+                                            vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                               vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                               vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                             vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                             vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                             vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tumu(vbool2_t vm, vbfloat16m8_t vd,
+                                             vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                            vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                            vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                          vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                          vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                          vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_mu(vbool2_t vm, vbfloat16m8_t vd,
+                                          vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_x_v_bf16m8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                           vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                           vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                           vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_mu(vbool2_t vm, vbfloat16m8_t vd,
+                                           vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_xu_v_bf16m8_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf4_f32mf2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16mf2_f32m1_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m1_f32m2_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m2_f32m4_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_f_v_bf16m4_f32m8_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmacc.c
new file mode 100644
index 0000000000000..f67b1001a5fcb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmacc.c
@@ -0,0 +1,1017 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_vv_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_vf_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmsac.c
new file mode 100644
index 0000000000000..6d78c74e14694
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmsac.c
@@ -0,0 +1,1017 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_vv_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_vf_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                               __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmul.c
new file mode 100644
index 0000000000000..9fcfe818be71f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwmul.c
@@ -0,0 +1,1015 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, vs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, rs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m1_f32m2_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m1_f32m2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m2_f32m4_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m2_f32m4_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_vv_bf16m4_f32m8_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_vf_bf16m4_f32m8_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmacc.c
new file mode 100644
index 0000000000000..73cc82219b201
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmacc.c
@@ -0,0 +1,1034 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    __bf16 vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                      vfloat32mf2_t vd,
+                                                      vbfloat16mf4_t vs1,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd,
+                                                    vbfloat16mf2_t vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd, __bf16 vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd,
+                                                   vbfloat16m1_t vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd, __bf16 vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   vbfloat16m2_t vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   vbfloat16m4_t vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_mu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_vv_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_vf_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmsac.c
new file mode 100644
index 0000000000000..6133230e5cd84
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwnmsac.c
@@ -0,0 +1,1034 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    __bf16 vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                  vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                 vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm_tu(vd, vs1, vs2, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm_tum(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                      vfloat32mf2_t vd,
+                                                      vbfloat16mf4_t vs1,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs1, vs2,
+                                                    __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd,
+                                                    vbfloat16mf2_t vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd, __bf16 vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs1, vs2,
+                                                   __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd,
+                                                   vbfloat16m1_t vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd, __bf16 vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   vbfloat16m2_t vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   vbfloat16m4_t vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_mu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs1, vs2,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs1, vs2,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m1_f32m2_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m2_f32m4_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_vv_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_vf_bf16m4_f32m8_rm_mu(vm, vd, vs1, vs2,
+                                                __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwsub.c
new file mode 100644
index 0000000000000..9d9b0b0aa2f3c
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vfwsub.c
@@ -0,0 +1,2007 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tu(vfloat32mf2_t vd, vfloat32mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             vfloat32mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           vfloat32m1_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           vfloat32m2_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           vfloat32m4_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           vfloat32m8_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              vfloat32mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            vfloat32m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                                vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                               vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm_tu(vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vfloat32mf2_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm_tum(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm_tum(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm_tum(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vfloat32m1_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm_tum(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vfloat32m2_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vfloat32m4_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm_tum(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vfloat32m8_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm_tum(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, vs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, rs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vfloat32mf2_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm_tumu(vm, vd, vs2, vs1,
+                                                  __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm_tumu(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, rs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vfloat32m1_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm_tumu(vm, vd, vs2, vs1,
+                                                 __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vfloat32m2_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vfloat32m4_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm_tumu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vfloat32m8_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm_tumu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, rs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf4_f32mf2_rm_mu(vm, vd, vs2, vs1,
+                                                __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32mf2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                             vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16mf2_f32m1_rm_mu(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_bf16mf2_f32m1_rm_mu(vm, vd, vs2, rs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16mf2_f32m1_rm_mu(vm, vd, vs2, vs1,
+                                               __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m1_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m1_f32m2_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m1_f32m2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m1_f32m2_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m2_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m2_f32m4_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m2_f32m4_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m2_f32m4_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m4_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_bf16m4_f32m8_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_bf16m4_f32m8_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_bf16m4_f32m8_rm_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE,
+                                              vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_bf16_f32m8_rm_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE,
+                                            vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfeq.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfeq.c
new file mode 100644
index 0000000000000..b96aae5615db8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfeq.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfge.c
new file mode 100644
index 0000000000000..47d0427abf99b
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfge.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfgt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfgt.c
new file mode 100644
index 0000000000000..0a0ead22361e9
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfgt.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfle.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfle.c
new file mode 100644
index 0000000000000..27ddefec8c9c0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfle.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmflt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmflt.c
new file mode 100644
index 0000000000000..d5f4f777580d3
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmflt.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfne.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfne.c
new file mode 100644
index 0000000000000..c2df9474acc72
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/non-overloaded/vmfne.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf4_b64_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16mf2_b32_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m1_b16_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m2_b8_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m4_b4_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne_vv_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_vf_bf16m8_b2_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfadd.c
new file mode 100644
index 0000000000000..2bd3b3995c6c8
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfadd_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfclass.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfclass.c
new file mode 100644
index 0000000000000..e2a993aca7069
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfclass.c
@@ -0,0 +1,272 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tu(vuint16mf4_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tu(vuint16mf2_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tu(vuint16m1_t vd, vbfloat16m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tu(vuint16m2_t vd, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tu(vuint16m4_t vd, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tu(vuint16m8_t vd, vbfloat16m8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfclass_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tum(vbool64_t vm, vuint16mf4_t vd,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tum(vbool32_t vm, vuint16mf2_t vd,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tum(vbool16_t vm, vuint16m1_t vd,
+                                            vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tum(vbool8_t vm, vuint16m2_t vd,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tum(vbool4_t vm, vuint16m4_t vd,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tum(vbool2_t vm, vuint16m8_t vd,
+                                            vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_tumu(vbool64_t vm, vuint16mf4_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_tumu(vbool32_t vm, vuint16mf2_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_tumu(vbool16_t vm, vuint16m1_t vd,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_tumu(vbool8_t vm, vuint16m2_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_tumu(vbool4_t vm, vuint16m4_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_tumu(vbool2_t vm, vuint16m8_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i16> @test_vfclass_v_bf16mf4_u16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i16> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1bf16.i64(<vscale x 1 x i16> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
+//
+vuint16mf4_t test_vfclass_v_bf16mf4_u16mf4_mu(vbool64_t vm, vuint16mf4_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i16> @test_vfclass_v_bf16mf2_u16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i16> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2bf16.i64(<vscale x 2 x i16> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
+//
+vuint16mf2_t test_vfclass_v_bf16mf2_u16mf2_mu(vbool32_t vm, vuint16mf2_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i16> @test_vfclass_v_bf16m1_u16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i16> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4bf16.i64(<vscale x 4 x i16> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
+//
+vuint16m1_t test_vfclass_v_bf16m1_u16m1_mu(vbool16_t vm, vuint16m1_t vd,
+                                           vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vfclass_v_bf16m2_u16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8bf16.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vfclass_v_bf16m2_u16m2_mu(vbool8_t vm, vuint16m2_t vd,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vfclass_v_bf16m4_u16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16bf16.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vfclass_v_bf16m4_u16m4_mu(vbool4_t vm, vuint16m4_t vd,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vfclass_v_bf16m8_u16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32bf16.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vfclass_v_bf16m8_u16m8_mu(vbool2_t vm, vuint16m8_t vd,
+                                           vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfclass_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmacc.c
new file mode 100644
index 0000000000000..eb7427107c942
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmacc.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmacc_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmacc_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmacc_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmacc_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmacc_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmacc_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmacc_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmacc_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmacc_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmacc_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmacc_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmacc_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmadd.c
new file mode 100644
index 0000000000000..68d490d04ff86
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmax.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmax.c
new file mode 100644
index 0000000000000..5f682e80b61b5
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmax.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmax_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmax_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmax_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmax_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmax_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmax_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmax_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmax_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmax_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmax_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmax_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmax_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmax_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmerge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmerge.c
new file mode 100644
index 0000000000000..9593ad5b5a592
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmerge.c
@@ -0,0 +1,69 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmerge_vfm_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmerge_vfm_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, vbool64_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmerge_vfm_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmerge_vfm_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, vbool32_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmerge_vfm_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmerge_vfm_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, vbool16_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmerge_vfm_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmerge_vfm_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, vbool8_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmerge_vfm_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmerge_vfm_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, vbool4_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmerge_vfm_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmerge_vfm_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, vbool2_t mask, size_t vl) {
+  return __riscv_vfmerge_tu(maskedoff, op1, op2, mask, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmin.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmin.c
new file mode 100644
index 0000000000000..f3ef3c3614027
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmin.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmin_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmin_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmin_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmin_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmin_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmin_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmin_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmin_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmin_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmin_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmin_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmin_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmin_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsac.c
new file mode 100644
index 0000000000000..0587c57af2b06
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsac.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsac_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsac_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsac_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsac_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsac_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsac_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsac_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsac_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsac_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsac_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsac_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsac_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsub.c
new file mode 100644
index 0000000000000..2ad26f8b24e12
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmul.c
new file mode 100644
index 0000000000000..d1e726a9d63ea
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmul.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmul_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmul_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmul_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmul_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmul_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmul_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmul_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmul_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmul_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmul_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmul_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmul_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfmul_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmv.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmv.c
new file mode 100644
index 0000000000000..9fd1ffce446bb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfmv.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_v_f_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_v_f_bf16mf4_tu(vbfloat16mf4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_v_f_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_v_f_bf16mf2_tu(vbfloat16mf2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_v_f_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_v_f_bf16m1_tu(vbfloat16m1_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_v_f_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_v_f_bf16m2_tu(vbfloat16m2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_v_f_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_v_f_bf16m4_tu(vbfloat16m4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_v_f_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_v_f_bf16m8_tu(vbfloat16m8_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_v_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfmv_s_f_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfmv_s_f_bf16mf4_tu(vbfloat16mf4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfmv_s_f_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfmv_s_f_bf16mf2_tu(vbfloat16mf2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfmv_s_f_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfmv_s_f_bf16m1_tu(vbfloat16m1_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfmv_s_f_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfmv_s_f_bf16m2_tu(vbfloat16m2_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfmv_s_f_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfmv_s_f_bf16m4_tu(vbfloat16m4_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfmv_s_f_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], bfloat noundef [[SRC:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], bfloat [[SRC]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfmv_s_f_bf16m8_tu(vbfloat16m8_t maskedoff, __bf16 src, size_t vl) {
+  return __riscv_vfmv_s_tu(maskedoff, src, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt.c
new file mode 100644
index 0000000000000..c6cd0a55fa530
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt.c
@@ -0,0 +1,1539 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tu(vint8mf8_t vd, vbfloat16mf4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tu(vint8mf4_t vd, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tu(vint8mf2_t vd, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                           size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tu(vuint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tu(vuint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tu(vuint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tu(vuint8m1_t vd, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tu(vuint8m2_t vd, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tu(vuint8m4_t vd, vbfloat16m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tu(vbfloat16mf4_t vd,
+                                            vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tu(vbfloat16mf2_t vd, vfloat32m1_t vs2,
+                                            size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tu(vbfloat16m1_t vd, vfloat32m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tu(vbfloat16m2_t vd, vfloat32m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tu(vbfloat16m4_t vd, vfloat32m8_t vs2,
+                                          size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tum(vbool64_t vm, vint8mf8_t vd,
+                                               vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tum(vbool32_t vm, vint8mf4_t vd,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tum(vbool16_t vm, vint8mf2_t vd,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tum(vbool8_t vm, vint8m1_t vd,
+                                            vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tum(vbool4_t vm, vint8m2_t vd,
+                                            vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tum(vbool2_t vm, vint8m4_t vd,
+                                            vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tum(vbool64_t vm, vuint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tum(vbool32_t vm, vuint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tum(vbool16_t vm, vuint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tum(vbool8_t vm, vuint8m1_t vd,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tum(vbool4_t vm, vuint8m2_t vd,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tum(vbool2_t vm, vuint8m4_t vd,
+                                              vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                           vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                           vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                           vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_tumu(vbool16_t vm, vint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_tumu(vbool8_t vm, vint8m1_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_tumu(vbool4_t vm, vint8m2_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_tumu(vbool2_t vm, vint8m4_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_tumu(vbool64_t vm, vuint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_tumu(vbool32_t vm, vuint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_tumu(vbool16_t vm, vuint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_tumu(vbool8_t vm, vuint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_tumu(vbool4_t vm, vuint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_tumu(vbool2_t vm, vuint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                            vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                            vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                            vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_mu(vbool64_t vm, vint8mf8_t vd,
+                                              vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_mu(vbool32_t vm, vint8mf4_t vd,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_mu(vbool16_t vm, vint8mf2_t vd,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_mu(vbool8_t vm, vint8m1_t vd,
+                                           vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_mu(vbool4_t vm, vint8m2_t vd,
+                                           vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_mu(vbool2_t vm, vint8m4_t vd,
+                                           vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_mu(vbool64_t vm, vuint8mf8_t vd,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_mu(vbool32_t vm, vuint8mf4_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_mu(vbool16_t vm, vuint8mf2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_mu(vbool8_t vm, vuint8m1_t vd,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_mu(vbool4_t vm, vuint8m2_t vd,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_mu(vbool2_t vm, vuint8m4_t vd,
+                                             vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                            vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                            vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                          vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                          vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                          vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tu(vint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tu(vint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tu(vint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                              size_t vl) {
+  return __riscv_vfncvt_x_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tu(vuint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tu(vuint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tu(vuint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tu(vuint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tu(vuint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tu(vuint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tu(vbfloat16mf4_t vd,
+                                               vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tu(vbfloat16mf2_t vd,
+                                               vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tu(vbfloat16m1_t vd, vfloat32m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tu(vbfloat16m2_t vd, vfloat32m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tu(vbfloat16m4_t vd, vfloat32m8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfncvt_f_bf16_tu(vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tum(vbool64_t vm, vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tum(vbool32_t vm, vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tum(vbool16_t vm, vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tum(vbool8_t vm, vint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tum(vbool4_t vm, vint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tum(vbool2_t vm, vint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tum(vbool64_t vm,
+                                                    vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tum(vbool32_t vm,
+                                                    vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tum(vbool16_t vm, vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tum(vbool8_t vm, vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tum(vbool4_t vm, vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tum(vbool2_t vm, vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tum(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_tumu(vbool16_t vm, vint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_tumu(vbool8_t vm, vint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_tumu(vbool4_t vm, vint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_tumu(vbool2_t vm, vint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_tumu(vbool64_t vm,
+                                                     vuint8mf8_t vd,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_tumu(vbool32_t vm,
+                                                     vuint8mf4_t vd,
+                                                     vbfloat16mf2_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_tumu(vbool16_t vm,
+                                                    vuint8mf2_t vd,
+                                                    vbfloat16m1_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_tumu(vbool8_t vm, vuint8m1_t vd,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_tumu(vbool4_t vm, vuint8m2_t vd,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_tumu(vbool2_t vm, vuint8m4_t vd,
+                                                  vbfloat16m8_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_tumu(vbool64_t vm,
+                                                 vbfloat16mf4_t vd,
+                                                 vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_tumu(vbool32_t vm,
+                                                 vbfloat16mf2_t vd,
+                                                 vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                               vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                               vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                               vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_tumu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_x_f_w_bf16mf4_i8mf8_rm_mu(vbool64_t vm, vint8mf8_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_x_f_w_bf16mf2_i8mf4_rm_mu(vbool32_t vm, vint8mf4_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_x_f_w_bf16m1_i8mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_x_f_w_bf16m1_i8mf2_rm_mu(vbool16_t vm, vint8mf2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_x_f_w_bf16m2_i8m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_x_f_w_bf16m2_i8m1_rm_mu(vbool8_t vm, vint8m1_t vd,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_x_f_w_bf16m4_i8m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_x_f_w_bf16m4_i8m2_rm_mu(vbool4_t vm, vint8m2_t vd,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_x_f_w_bf16m8_i8m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_x_f_w_bf16m8_i8m4_rm_mu(vbool2_t vm, vint8m4_t vd,
+                                              vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_x_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_xu_f_w_bf16mf4_u8mf8_rm_mu(vbool64_t vm, vuint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_xu_f_w_bf16mf2_u8mf4_rm_mu(vbool32_t vm, vuint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_xu_f_w_bf16m1_u8mf2_rm_mu(vbool16_t vm, vuint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_xu_f_w_bf16m2_u8m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_xu_f_w_bf16m2_u8m1_rm_mu(vbool8_t vm, vuint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_xu_f_w_bf16m4_u8m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_xu_f_w_bf16m4_u8m2_rm_mu(vbool4_t vm, vuint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_xu_f_w_bf16m8_u8m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_xu_f_w_bf16m8_u8m4_rm_mu(vbool2_t vm, vuint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_xu_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_f_f_w_bf16mf4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_f_f_w_bf16mf4_rm_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                               vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_f_f_w_bf16mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_f_f_w_bf16mf2_rm_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                               vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_f_f_w_bf16m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_f_f_w_bf16m1_rm_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                             vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_f_f_w_bf16m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_f_f_w_bf16m2_rm_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                             vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_f_f_w_bf16m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_f_f_w_bf16m4_rm_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                             vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_f_bf16_mu(vm, vd, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rod.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rod.c
new file mode 100644
index 0000000000000..0745633042d44
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rod.c
@@ -0,0 +1,233 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tu(vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tu(vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tu(vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tu(vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tu(vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tum(vbool64_t vm,
+                                                 vbfloat16mf4_t vd,
+                                                 vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tum(vbool32_t vm,
+                                                 vbfloat16mf2_t vd,
+                                                 vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                               vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                               vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                               vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_tumu(vbool64_t vm,
+                                                  vbfloat16mf4_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_tumu(vbool32_t vm,
+                                                  vbfloat16mf2_t vd,
+                                                  vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                                vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                                vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                                vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfncvt_rod_f_f_w_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                                vfloat32mf2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfncvt_rod_f_f_w_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfncvt_rod_f_f_w_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                                vfloat32m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfncvt_rod_f_f_w_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfncvt_rod_f_f_w_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                              vfloat32m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfncvt_rod_f_f_w_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfncvt_rod_f_f_w_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                              vfloat32m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfncvt_rod_f_f_w_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfncvt_rod_f_f_w_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                              vfloat32m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rod_f_bf16_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rtz.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rtz.c
new file mode 100644
index 0000000000000..b906c5f411064
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfncvt_rtz.c
@@ -0,0 +1,572 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tu(vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tu(vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tu(vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tu(vint8m1_t vd, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tu(vint8m2_t vd, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tu(vint8m4_t vd, vbfloat16m8_t vs2,
+                                               size_t vl) {
+  return __riscv_vfncvt_rtz_x_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tu(
+// CHECK-RV64-SAME: <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tu(vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tu(
+// CHECK-RV64-SAME: <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tu(vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tu(vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tu(
+// CHECK-RV64-SAME: <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tu(vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tu(
+// CHECK-RV64-SAME: <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tu(vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tu(
+// CHECK-RV64-SAME: <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tu(vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tum(vbool64_t vm, vint8mf8_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tum(vbool32_t vm, vint8mf4_t vd,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tum(vbool16_t vm, vint8mf2_t vd,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tum(vbool8_t vm, vint8m1_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tum(vbool4_t vm, vint8m2_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tum(vbool2_t vm, vint8m4_t vd,
+                                                vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tum(vbool64_t vm,
+                                                     vuint8mf8_t vd,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tum(vbool32_t vm,
+                                                     vuint8mf4_t vd,
+                                                     vbfloat16mf2_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tum(vbool16_t vm,
+                                                    vuint8mf2_t vd,
+                                                    vbfloat16m1_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tum(vbool8_t vm, vuint8m1_t vd,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tum(vbool4_t vm, vuint8m2_t vd,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tum(vbool2_t vm, vuint8m4_t vd,
+                                                  vbfloat16m8_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_tumu(vbool64_t vm, vint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_tumu(vbool32_t vm, vint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_tumu(vbool16_t vm, vint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_tumu(vbool8_t vm, vint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_tumu(vbool4_t vm, vint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_tumu(vbool2_t vm, vint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_tumu(vbool64_t vm,
+                                                      vuint8mf8_t vd,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_tumu(vbool32_t vm,
+                                                      vuint8mf4_t vd,
+                                                      vbfloat16mf2_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_tumu(vbool16_t vm,
+                                                     vuint8mf2_t vd,
+                                                     vbfloat16m1_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_tumu(vbool8_t vm, vuint8m1_t vd,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_tumu(vbool4_t vm, vuint8m2_t vd,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_tumu(vbool2_t vm, vuint8m4_t vd,
+                                                   vbfloat16m8_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vint8mf8_t test_vfncvt_rtz_x_f_w_bf16mf4_i8mf8_mu(vbool64_t vm, vint8mf8_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vint8mf4_t test_vfncvt_rtz_x_f_w_bf16mf2_i8mf4_mu(vbool32_t vm, vint8mf4_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vint8mf2_t test_vfncvt_rtz_x_f_w_bf16m1_i8mf2_mu(vbool16_t vm, vint8mf2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_x_f_w_bf16m2_i8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vint8m1_t test_vfncvt_rtz_x_f_w_bf16m2_i8m1_mu(vbool8_t vm, vint8m1_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_x_f_w_bf16m4_i8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vint8m2_t test_vfncvt_rtz_x_f_w_bf16m4_i8m2_mu(vbool4_t vm, vint8m2_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_x_f_w_bf16m8_i8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vint8m4_t test_vfncvt_rtz_x_f_w_bf16m8_i8m4_mu(vbool2_t vm, vint8m4_t vd,
+                                               vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_x_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x i8> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16.i64(<vscale x 1 x i8> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
+//
+vuint8mf8_t test_vfncvt_rtz_xu_f_w_bf16mf4_u8mf8_mu(vbool64_t vm,
+                                                    vuint8mf8_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i8> @test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x i8> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16.i64(<vscale x 2 x i8> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
+//
+vuint8mf4_t test_vfncvt_rtz_xu_f_w_bf16mf2_u8mf4_mu(vbool32_t vm,
+                                                    vuint8mf4_t vd,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i8> @test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x i8> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16.i64(<vscale x 4 x i8> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
+//
+vuint8mf2_t test_vfncvt_rtz_xu_f_w_bf16m1_u8mf2_mu(vbool16_t vm, vuint8mf2_t vd,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i8> @test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x i8> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16.i64(<vscale x 8 x i8> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
+//
+vuint8m1_t test_vfncvt_rtz_xu_f_w_bf16m2_u8m1_mu(vbool8_t vm, vuint8m1_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i8> @test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x i8> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16.i64(<vscale x 16 x i8> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
+//
+vuint8m2_t test_vfncvt_rtz_xu_f_w_bf16m4_u8m2_mu(vbool4_t vm, vuint8m2_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i8> @test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x i8> [[VD:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16.i64(<vscale x 32 x i8> [[VD]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
+//
+vuint8m4_t test_vfncvt_rtz_xu_f_w_bf16m8_u8m4_mu(vbool2_t vm, vuint8m4_t vd,
+                                                 vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfncvt_rtz_xu_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmacc.c
new file mode 100644
index 0000000000000..cc487b49429dd
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmacc.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmacc_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmacc_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmacc_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmacc_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmacc_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmacc_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmacc_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmacc_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmacc_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmacc_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmacc_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmacc.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmacc_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmacc_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmadd.c
new file mode 100644
index 0000000000000..f9c348b3dbb0b
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmadd.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmadd_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmadd_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmadd_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmadd_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmadd_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmadd_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmadd_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmadd_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmadd_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmadd_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmadd_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmadd.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmadd_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmadd_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsac.c
new file mode 100644
index 0000000000000..83d35e81403ce
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsac.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsac_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsac_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsac_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsac_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsac_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsac_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsac_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsac_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsac_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsac_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsac_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsac.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsac_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsac_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsub.c
new file mode 100644
index 0000000000000..f5282a195131d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfnmsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tu(vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tu(vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tu(vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tu(vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tu(vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tu(vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tu(vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tu(vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tu(vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tu(vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tu(vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tu(vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tu(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tum(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_tumu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, vbfloat16mf4_t vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfnmsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfnmsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t vd, __bf16 rs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, vbfloat16mf2_t vs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfnmsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfnmsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t vd, __bf16 rs1, vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, vbfloat16m1_t vs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfnmsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfnmsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t vd, __bf16 rs1, vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, vbfloat16m2_t vs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfnmsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfnmsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t vd, __bf16 rs1, vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, vbfloat16m4_t vs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfnmsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfnmsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t vd, __bf16 rs1, vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x bfloat> [[VS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x bfloat> [[VS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, vbfloat16m8_t vs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfnmsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], bfloat noundef [[RS1:%.*]], <vscale x 32 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfnmsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[VD]], bfloat [[RS1]], <vscale x 32 x bfloat> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfnmsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t vd, __bf16 rs1, vbfloat16m8_t vs2, size_t vl) {
+  return __riscv_vfnmsub_mu(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrec7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrec7.c
new file mode 100644
index 0000000000000..f8e5a339c87d7
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrec7.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrec7_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrec7_v_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrec7_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrec7_v_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrec7_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrec7_v_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrec7_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrec7_v_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrec7_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrec7_v_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrec7_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrec7_v_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrec7_mu(mask, maskedoff, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsqrt7.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsqrt7.c
new file mode 100644
index 0000000000000..7c6c926e50c10
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsqrt7.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tu(maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tum(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_tumu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsqrt7_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsqrt7_v_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsqrt7_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsqrt7_v_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsqrt7_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsqrt7_v_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsqrt7_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsqrt7_v_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsqrt7_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsqrt7_v_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsqrt7_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsqrt7_v_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, size_t vl) {
+  return __riscv_vfrsqrt7_mu(mask, maskedoff, op1, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsub.c
new file mode 100644
index 0000000000000..c09caeb8207af
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfrsub.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfrsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfrsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfrsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfrsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfrsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfrsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfrsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfrsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfrsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfrsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfrsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfrsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfrsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnj.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnj.c
new file mode 100644
index 0000000000000..c1f69932f6a02
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnj.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnj_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnj_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnj_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnj_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnj_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnj_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnj_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnj_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnj_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnj_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnj_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnj_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnj_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjn.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjn.c
new file mode 100644
index 0000000000000..1b799d87d8131
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjn.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjn_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjn_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjn_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjn_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjn_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjn_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjn_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjn_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjn_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjn_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjn_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjn_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjn_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjx.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjx.c
new file mode 100644
index 0000000000000..9c5f2af3d6e8f
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsgnjx.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsgnjx_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsgnjx_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsgnjx_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsgnjx_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsgnjx_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsgnjx_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsgnjx_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsgnjx_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsgnjx_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsgnjx_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsgnjx_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsgnjx_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsgnjx_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1down.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1down.c
new file mode 100644
index 0000000000000..691302e245427
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1down.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1down_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1down_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1down_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1down_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1down_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1down_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1down_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1down_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1down_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1down_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1down_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1down_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1down_mu(mask, maskedoff, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1up.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1up.c
new file mode 100644
index 0000000000000..1238d2204c6d4
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfslide1up.c
@@ -0,0 +1,249 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tu(maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tum(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_tumu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfslide1up_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 1 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfslide1up_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfslide1up_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 2 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfslide1up_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfslide1up_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 4 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfslide1up_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfslide1up_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 8 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfslide1up_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfslide1up_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 16 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfslide1up_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfslide1up_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[SRC:%.*]], bfloat noundef [[VALUE:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[SRC]], bfloat [[VALUE]], <vscale x 32 x i1> [[MASK]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfslide1up_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t src, __bf16 value, size_t vl) {
+  return __riscv_vfslide1up_mu(mask, maskedoff, src, value, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsub.c
new file mode 100644
index 0000000000000..ea4f8f043d6bb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfsub.c
@@ -0,0 +1,489 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tu(vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tu(vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tu(vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tu(vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tu(vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tu(vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tu(maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tum(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tum(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tum(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tum(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tum(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tum(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tum(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_tumu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_tumu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_tumu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_tumu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_tumu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_tumu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_tumu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vv_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vv_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfsub_vf_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x bfloat> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16.i64(<vscale x 1 x bfloat> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfsub_vf_bf16mf4_mu(vbool64_t mask, vbfloat16mf4_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vv_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vv_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfsub_vf_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x bfloat> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16.i64(<vscale x 2 x bfloat> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfsub_vf_bf16mf2_mu(vbool32_t mask, vbfloat16mf2_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vv_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vv_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfsub_vf_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x bfloat> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16.i64(<vscale x 4 x bfloat> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfsub_vf_bf16m1_mu(vbool16_t mask, vbfloat16m1_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vv_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vv_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfsub_vf_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x bfloat> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16.i64(<vscale x 8 x bfloat> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfsub_vf_bf16m2_mu(vbool8_t mask, vbfloat16m2_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vv_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vv_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfsub_vf_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x bfloat> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16.i64(<vscale x 16 x bfloat> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfsub_vf_bf16m4_mu(vbool4_t mask, vbfloat16m4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vv_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vv_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfsub_vf_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x bfloat> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16.i64(<vscale x 32 x bfloat> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfsub_vf_bf16m8_mu(vbool2_t mask, vbfloat16m8_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vfsub_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwadd.c
new file mode 100644
index 0000000000000..e5b7b8da1f3cc
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwadd.c
@@ -0,0 +1,1932 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tu(vfloat32mf2_t vd, vfloat32mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             vfloat32mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           vfloat32m1_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           vfloat32m2_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           vfloat32m4_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           vfloat32m8_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              vfloat32mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            vfloat32m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vfloat32mf2_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vfloat32m1_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vfloat32m2_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vfloat32m4_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vfloat32m8_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwadd_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vfloat32mf2_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vfloat32m1_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vfloat32m2_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vfloat32m4_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwadd_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vfloat32m8_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwadd_wf_bf16_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwadd_wf_bf16_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwadd_wf_bf16_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwadd_wf_bf16_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwadd_wf_bf16_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwadd_wf_bf16_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwadd_wf_bf16_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwadd_wf_bf16_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwadd_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwadd_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwadd_wf_bf16_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwadd_wf_bf16_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwadd_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwcvt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwcvt.c
new file mode 100644
index 0000000000000..730010421b944
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwcvt.c
@@ -0,0 +1,765 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tu(vbfloat16mf4_t vd, vint8mf8_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tu(vbfloat16mf2_t vd, vint8mf4_t vs2,
+                                            size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tu(vbfloat16m1_t vd, vint8mf2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tu(vbfloat16m2_t vd, vint8m1_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tu(vbfloat16m4_t vd, vint8m2_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tu(vbfloat16m8_t vd, vint8m4_t vs2,
+                                          size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tu(
+// CHECK-RV64-SAME: <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tu(vbfloat16mf4_t vd, vuint8mf8_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tu(
+// CHECK-RV64-SAME: <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tu(vbfloat16mf2_t vd, vuint8mf4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tu(
+// CHECK-RV64-SAME: <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tu(vbfloat16m1_t vd, vuint8mf2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tu(
+// CHECK-RV64-SAME: <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tu(vbfloat16m2_t vd, vuint8m1_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tu(
+// CHECK-RV64-SAME: <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tu(vbfloat16m4_t vd, vuint8m2_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tu(
+// CHECK-RV64-SAME: <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tu(vbfloat16m8_t vd, vuint8m4_t vs2,
+                                           size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tu(vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                           vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                           vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                           vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tum(vbool2_t vm, vbfloat16m8_t vd,
+                                           vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tum(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tum(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tum(vbool16_t vm, vbfloat16m1_t vd,
+                                            vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tum(vbool8_t vm, vbfloat16m2_t vd,
+                                            vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tum(vbool4_t vm, vbfloat16m4_t vd,
+                                            vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tum(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tum(vbool2_t vm, vbfloat16m8_t vd,
+                                            vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwcvt_f_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwcvt_f_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tum(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                              vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                              vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                            vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                            vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                            vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_tumu(vbool2_t vm, vbfloat16m8_t vd,
+                                            vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_tumu(vbool64_t vm, vbfloat16mf4_t vd,
+                                               vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_tumu(vbool32_t vm, vbfloat16mf2_t vd,
+                                               vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_tumu(vbool16_t vm, vbfloat16m1_t vd,
+                                             vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_tumu(vbool8_t vm, vbfloat16m2_t vd,
+                                             vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_tumu(vbool4_t vm, vbfloat16m4_t vd,
+                                             vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_tumu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_tumu(vbool2_t vm, vbfloat16m8_t vd,
+                                             vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwcvt_f_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_tumu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_x_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_x_v_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                            vint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_x_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_x_v_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                            vint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_x_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_x_v_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                          vint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_x_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_x_v_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                          vint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_x_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_x_v_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                          vint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_x_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_x_v_bf16m8_mu(vbool2_t vm, vbfloat16m8_t vd,
+                                          vint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x bfloat> @test_vfwcvt_f_xu_v_bf16mf4_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x bfloat> [[VD:%.*]], <vscale x 1 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8.i64(<vscale x 1 x bfloat> [[VD]], <vscale x 1 x i8> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x bfloat> [[TMP0]]
+//
+vbfloat16mf4_t test_vfwcvt_f_xu_v_bf16mf4_mu(vbool64_t vm, vbfloat16mf4_t vd,
+                                             vuint8mf8_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x bfloat> @test_vfwcvt_f_xu_v_bf16mf2_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x bfloat> [[VD:%.*]], <vscale x 2 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8.i64(<vscale x 2 x bfloat> [[VD]], <vscale x 2 x i8> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x bfloat> [[TMP0]]
+//
+vbfloat16mf2_t test_vfwcvt_f_xu_v_bf16mf2_mu(vbool32_t vm, vbfloat16mf2_t vd,
+                                             vuint8mf4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x bfloat> @test_vfwcvt_f_xu_v_bf16m1_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x bfloat> [[VD:%.*]], <vscale x 4 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8.i64(<vscale x 4 x bfloat> [[VD]], <vscale x 4 x i8> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x bfloat> [[TMP0]]
+//
+vbfloat16m1_t test_vfwcvt_f_xu_v_bf16m1_mu(vbool16_t vm, vbfloat16m1_t vd,
+                                           vuint8mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x bfloat> @test_vfwcvt_f_xu_v_bf16m2_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x bfloat> [[VD:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8.i64(<vscale x 8 x bfloat> [[VD]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x bfloat> [[TMP0]]
+//
+vbfloat16m2_t test_vfwcvt_f_xu_v_bf16m2_mu(vbool8_t vm, vbfloat16m2_t vd,
+                                           vuint8m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x bfloat> @test_vfwcvt_f_xu_v_bf16m4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x bfloat> [[VD:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8.i64(<vscale x 16 x bfloat> [[VD]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x bfloat> [[TMP0]]
+//
+vbfloat16m4_t test_vfwcvt_f_xu_v_bf16m4_mu(vbool4_t vm, vbfloat16m4_t vd,
+                                           vuint8m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x bfloat> @test_vfwcvt_f_xu_v_bf16m8_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[VM:%.*]], <vscale x 32 x bfloat> [[VD:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8.i64(<vscale x 32 x bfloat> [[VD]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 32 x bfloat> [[TMP0]]
+//
+vbfloat16m8_t test_vfwcvt_f_xu_v_bf16m8_mu(vbool2_t vm, vbfloat16m8_t vd,
+                                           vuint8m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_bf16_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwcvt_f_f_v_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwcvt_f_f_v_bf16mf4_f32mf2_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwcvt_f_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwcvt_f_f_v_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwcvt_f_f_v_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwcvt_f_f_v_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwcvt_f_f_v_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwcvt_f_f_v_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwcvt_f_f_v_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_mu(vm, vd, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwcvt_f_f_v_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwcvt.f.f.v.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwcvt_f_f_v_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwcvt_f_mu(vm, vd, vs2, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmacc.c
new file mode 100644
index 0000000000000..b05f88028874d
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmacc.c
@@ -0,0 +1,977 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmacc_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmacc_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmacc_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmacc_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmacc_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmacc_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmacc_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmacc_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmacc_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmacc_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmsac.c
new file mode 100644
index 0000000000000..93721f6a889d9
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmsac.c
@@ -0,0 +1,977 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs1,
+                                                vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                __bf16 vs1, vbfloat16mf4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs1,
+                                              vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                              __bf16 vs1, vbfloat16mf2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs1,
+                                             vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             __bf16 vs1, vbfloat16m1_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs1,
+                                             vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             __bf16 vs1, vbfloat16m2_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs1,
+                                             vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             __bf16 vs1, vbfloat16m4_t vs2,
+                                             size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmsac_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmsac_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmsac_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmsac_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmsac_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmsac_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmsac_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmsac_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmsac_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmsac_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmul.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmul.c
new file mode 100644
index 0000000000000..4a2b5e39ef2cf
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwmul.c
@@ -0,0 +1,975 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwmul_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwmul_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwmul_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwmul_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwmul_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwmul_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwmul_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwmul_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwmul_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwmul_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwmul_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwmul_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmacc.c
new file mode 100644
index 0000000000000..57e433441cd40
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmacc.c
@@ -0,0 +1,994 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    __bf16 vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                      vfloat32mf2_t vd,
+                                                      vbfloat16mf4_t vs1,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd,
+                                                    vbfloat16mf2_t vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd, __bf16 vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd,
+                                                   vbfloat16m1_t vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd, __bf16 vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   vbfloat16m2_t vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   vbfloat16m4_t vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmacc_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmacc_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmacc_vf_bf16mf4_f32mf2_rm_mu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmacc_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmacc_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmacc_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmacc_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmacc_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmacc_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmacc_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmacc_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmacc_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmsac.c
new file mode 100644
index 0000000000000..42da060126a2b
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwnmsac.c
@@ -0,0 +1,994 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd, __bf16 vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd, __bf16 vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tu(vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, __bf16 vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tu(vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, __bf16 vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tu(vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, __bf16 vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tum(vbool64_t vm,
+                                                  vfloat32mf2_t vd, __bf16 vs1,
+                                                  vbfloat16mf4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs1,
+                                                vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                __bf16 vs1, vbfloat16mf2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs1,
+                                               vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                               __bf16 vs1, vbfloat16m1_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs1,
+                                               vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                               __bf16 vs1, vbfloat16m2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs1,
+                                               vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                               __bf16 vs1, vbfloat16m4_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_tumu(vbool64_t vm,
+                                                   vfloat32mf2_t vd, __bf16 vs1,
+                                                   vbfloat16mf4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs1,
+                                                 vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                 __bf16 vs1, vbfloat16mf2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs1,
+                                                vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                __bf16 vs1, vbfloat16m1_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs1,
+                                                vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                __bf16 vs1, vbfloat16m2_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs1,
+                                                vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                __bf16 vs1, vbfloat16m4_t vs2,
+                                                size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs1,
+                                                 vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 __bf16 vs1, vbfloat16mf4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs1,
+                                               vbfloat16mf2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                               __bf16 vs1, vbfloat16mf2_t vs2,
+                                               size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs1,
+                                              vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                              __bf16 vs1, vbfloat16m1_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs1,
+                                              vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                              __bf16 vs1, vbfloat16m2_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs1,
+                                              vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                              __bf16 vs1, vbfloat16m4_t vs2,
+                                              size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                    __bf16 vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd, __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd, __bf16 vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd, __bf16 vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd, __bf16 vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tu(vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                     vfloat32mf2_t vd,
+                                                     vbfloat16mf4_t vs1,
+                                                     vbfloat16mf4_t vs2,
+                                                     size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tum(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd,
+                                                   vbfloat16mf2_t vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm,
+                                                   vfloat32m1_t vd, __bf16 vs1,
+                                                   vbfloat16mf2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  vbfloat16m1_t vs1,
+                                                  vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                  __bf16 vs1, vbfloat16m1_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  vbfloat16m2_t vs1,
+                                                  vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                  __bf16 vs1, vbfloat16m2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  vbfloat16m4_t vs1,
+                                                  vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                  __bf16 vs1, vbfloat16m4_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_tum(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                      vfloat32mf2_t vd,
+                                                      vbfloat16mf4_t vs1,
+                                                      vbfloat16mf4_t vs2,
+                                                      size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_tumu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd,
+                                                    vbfloat16mf2_t vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm,
+                                                    vfloat32m1_t vd, __bf16 vs1,
+                                                    vbfloat16mf2_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd,
+                                                   vbfloat16m1_t vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm,
+                                                   vfloat32m2_t vd, __bf16 vs1,
+                                                   vbfloat16m1_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   vbfloat16m2_t vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m2_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   vbfloat16m4_t vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                   __bf16 vs1,
+                                                   vbfloat16m4_t vs2,
+                                                   size_t vl) {
+  return __riscv_vfwnmsac_tumu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs1,
+                                                    vbfloat16mf4_t vs2,
+                                                    size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwnmsac_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], bfloat [[VS1]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwnmsac_vf_bf16mf4_f32mf2_rm_mu(
+    vbool64_t vm, vfloat32mf2_t vd, __bf16 vs1, vbfloat16mf4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwnmsac_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], bfloat [[VS1]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwnmsac_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                  __bf16 vs1,
+                                                  vbfloat16mf2_t vs2,
+                                                  size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs1,
+                                                 vbfloat16m1_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwnmsac_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], bfloat [[VS1]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwnmsac_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                                 __bf16 vs1, vbfloat16m1_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs1,
+                                                 vbfloat16m2_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwnmsac_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], bfloat [[VS1]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwnmsac_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                                 __bf16 vs1, vbfloat16m2_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs1,
+                                                 vbfloat16m4_t vs2, size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwnmsac_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], bfloat noundef [[VS1:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], bfloat [[VS1]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwnmsac_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                                 __bf16 vs1, vbfloat16m4_t vs2,
+                                                 size_t vl) {
+  return __riscv_vfwnmsac_mu(vm, vd, vs1, vs2, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwsub.c
new file mode 100644
index 0000000000000..1378bc963b216
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vfwsub.c
@@ -0,0 +1,1932 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tu(vfloat32mf2_t vd, vfloat32mf2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tu(vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tu(vfloat32m2_t vd, vbfloat16m1_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tu(vfloat32m4_t vd, vbfloat16m2_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tu(vfloat32m8_t vd, vbfloat16m4_t vs2,
+                                            __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 7, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                          __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vbfloat16mf4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2,
+                                                vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                             vfloat32mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vbfloat16mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2,
+                                              vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tum(vbool32_t vm, vfloat32m1_t vd,
+                                           vfloat32m1_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vbfloat16m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2,
+                                             vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tum(vbool16_t vm, vfloat32m2_t vd,
+                                           vfloat32m2_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vbfloat16m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2,
+                                             vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tum(vbool8_t vm, vfloat32m4_t vd,
+                                           vfloat32m4_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vbfloat16m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2,
+                                             vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tum(vbool4_t vm, vfloat32m8_t vd,
+                                           vfloat32m8_t vs2, __bf16 rs1,
+                                           size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vbfloat16mf4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2,
+                                                 vbfloat16mf4_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                              vfloat32mf2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vbfloat16mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2,
+                                               vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                            vfloat32m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vbfloat16m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2,
+                                              vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vbfloat16m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2,
+                                              vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vbfloat16m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2,
+                                              vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vbfloat16mf4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2,
+                                               vbfloat16mf4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                            vfloat32mf2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vbfloat16mf2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2,
+                                             vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_mu(vbool32_t vm, vfloat32m1_t vd,
+                                          vfloat32m1_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2,
+                                            vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vbfloat16m1_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                            vfloat32m2_t vs2, vbfloat16m1_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_mu(vbool16_t vm, vfloat32m2_t vd,
+                                          vfloat32m2_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2,
+                                            vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vbfloat16m2_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                            vfloat32m4_t vs2, vbfloat16m2_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_mu(vbool8_t vm, vfloat32m4_t vd,
+                                          vfloat32m4_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2,
+                                            vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vbfloat16m4_t vs2, __bf16 rs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                            vfloat32m8_t vs2, vbfloat16m4_t vs1,
+                                            size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 7, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_mu(vbool4_t vm, vfloat32m8_t vd,
+                                          vfloat32m8_t vs2, __bf16 rs1,
+                                          size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tu(vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tu(vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tu(
+// CHECK-RV64-SAME: <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tu(vfloat32m1_t vd, vfloat32m1_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tu(vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tu(
+// CHECK-RV64-SAME: <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tu(vfloat32m2_t vd, vfloat32m2_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tu(vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tu(
+// CHECK-RV64-SAME: <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tu(vfloat32m4_t vd, vfloat32m4_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tu(vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tu(vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tu(
+// CHECK-RV64-SAME: <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], i64 0, i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tu(vfloat32m8_t vd, vfloat32m8_t vs2,
+                                             __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_wf_tu(vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vbfloat16mf4_t vs2,
+                                                   __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tum(vbool64_t vm,
+                                                   vfloat32mf2_t vd,
+                                                   vfloat32mf2_t vs2,
+                                                   vbfloat16mf4_t vs1,
+                                                   size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tum(vbool64_t vm, vfloat32mf2_t vd,
+                                                vfloat32mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vbfloat16mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                                 vfloat32m1_t vs2,
+                                                 vbfloat16mf2_t vs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tum(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tum(vbool32_t vm, vfloat32m1_t vd,
+                                              vfloat32m1_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vbfloat16m1_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                                vfloat32m2_t vs2,
+                                                vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tum(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tum(vbool16_t vm, vfloat32m2_t vd,
+                                              vfloat32m2_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vbfloat16m2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                                vfloat32m4_t vs2,
+                                                vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tum(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tum(vbool8_t vm, vfloat32m4_t vd,
+                                              vfloat32m4_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vbfloat16m4_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                                vfloat32m8_t vs2,
+                                                vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tum(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tum(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 2)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tum(vbool4_t vm, vfloat32m8_t vd,
+                                              vfloat32m8_t vs2, __bf16 rs1,
+                                              size_t vl) {
+  return __riscv_vfwsub_wf_tum(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vbfloat16mf4_t vs2,
+                                                    __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_tumu(vbool64_t vm,
+                                                    vfloat32mf2_t vd,
+                                                    vfloat32mf2_t vs2,
+                                                    vbfloat16mf4_t vs1,
+                                                    size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_tumu(vbool64_t vm, vfloat32mf2_t vd,
+                                                 vfloat32mf2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vbfloat16mf2_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                                  vfloat32m1_t vs2,
+                                                  vbfloat16mf2_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_tumu(vbool32_t vm, vfloat32m1_t vd,
+                                               vfloat32m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vbfloat16m1_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                                 vfloat32m2_t vs2,
+                                                 vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_tumu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vbfloat16m2_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                                 vfloat32m4_t vs2,
+                                                 vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_tumu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vbfloat16m4_t vs2, __bf16 rs1,
+                                                 size_t vl) {
+  return __riscv_vfwsub_vf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                                 vfloat32m8_t vs2,
+                                                 vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_tumu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_tumu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 0)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_tumu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_tumu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_vf_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_vf_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vbfloat16mf4_t vs2,
+                                                  __bf16 rs1, size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wv_bf16mf4_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], <vscale x 1 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], <vscale x 1 x bfloat> [[VS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wv_bf16mf4_f32mf2_rm_mu(vbool64_t vm,
+                                                  vfloat32mf2_t vd,
+                                                  vfloat32mf2_t vs2,
+                                                  vbfloat16mf4_t vs1,
+                                                  size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x float> @test_vfwsub_wf_bf16_f32mf2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[VM:%.*]], <vscale x 1 x float> [[VD:%.*]], <vscale x 1 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16.i64(<vscale x 1 x float> [[VD]], <vscale x 1 x float> [[VS2]], bfloat [[RS1]], <vscale x 1 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
+//
+vfloat32mf2_t test_vfwsub_wf_bf16_f32mf2_rm_mu(vbool64_t vm, vfloat32mf2_t vd,
+                                               vfloat32mf2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_vf_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_vf_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vbfloat16mf2_t vs2, __bf16 rs1,
+                                                size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wv_bf16mf2_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], <vscale x 2 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], <vscale x 2 x bfloat> [[VS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wv_bf16mf2_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                                vfloat32m1_t vs2,
+                                                vbfloat16mf2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x float> @test_vfwsub_wf_bf16_f32m1_rm_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[VM:%.*]], <vscale x 2 x float> [[VD:%.*]], <vscale x 2 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16.i64(<vscale x 2 x float> [[VD]], <vscale x 2 x float> [[VS2]], bfloat [[RS1]], <vscale x 2 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
+//
+vfloat32m1_t test_vfwsub_wf_bf16_f32m1_rm_mu(vbool32_t vm, vfloat32m1_t vd,
+                                             vfloat32m1_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_vf_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_vf_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vbfloat16m1_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wv_bf16m1_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], <vscale x 4 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], <vscale x 4 x bfloat> [[VS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wv_bf16m1_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                               vfloat32m2_t vs2,
+                                               vbfloat16m1_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x float> @test_vfwsub_wf_bf16_f32m2_rm_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[VM:%.*]], <vscale x 4 x float> [[VD:%.*]], <vscale x 4 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16.i64(<vscale x 4 x float> [[VD]], <vscale x 4 x float> [[VS2]], bfloat [[RS1]], <vscale x 4 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
+//
+vfloat32m2_t test_vfwsub_wf_bf16_f32m2_rm_mu(vbool16_t vm, vfloat32m2_t vd,
+                                             vfloat32m2_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_vf_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_vf_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vbfloat16m2_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wv_bf16m2_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], <vscale x 8 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], <vscale x 8 x bfloat> [[VS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wv_bf16m2_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                               vfloat32m4_t vs2,
+                                               vbfloat16m2_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x float> @test_vfwsub_wf_bf16_f32m4_rm_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[VM:%.*]], <vscale x 8 x float> [[VD:%.*]], <vscale x 8 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16.i64(<vscale x 8 x float> [[VD]], <vscale x 8 x float> [[VS2]], bfloat [[RS1]], <vscale x 8 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
+//
+vfloat32m4_t test_vfwsub_wf_bf16_f32m4_rm_mu(vbool8_t vm, vfloat32m4_t vd,
+                                             vfloat32m4_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_vv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_vf_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x bfloat> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x bfloat> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_vf_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vbfloat16m4_t vs2, __bf16 rs1,
+                                               size_t vl) {
+  return __riscv_vfwsub_vf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wv_bf16m4_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], <vscale x 16 x bfloat> [[VS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], <vscale x 16 x bfloat> [[VS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wv_bf16m4_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                               vfloat32m8_t vs2,
+                                               vbfloat16m4_t vs1, size_t vl) {
+  return __riscv_vfwsub_wv_mu(vm, vd, vs2, vs1, __RISCV_FRM_RNE, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x float> @test_vfwsub_wf_bf16_f32m8_rm_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[VM:%.*]], <vscale x 16 x float> [[VD:%.*]], <vscale x 16 x float> [[VS2:%.*]], bfloat noundef [[RS1:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16.i64(<vscale x 16 x float> [[VD]], <vscale x 16 x float> [[VS2]], bfloat [[RS1]], <vscale x 16 x i1> [[VM]], i64 0, i64 [[VL]], i64 1)
+// CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
+//
+vfloat32m8_t test_vfwsub_wf_bf16_f32m8_rm_mu(vbool4_t vm, vfloat32m8_t vd,
+                                             vfloat32m8_t vs2, __bf16 rs1,
+                                             size_t vl) {
+  return __riscv_vfwsub_wf_mu(vm, vd, vs2, rs1, __RISCV_FRM_RNE, vl);
+}
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfeq.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfeq.c
new file mode 100644
index 0000000000000..3945f826809b0
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfeq.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfeq_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfeq_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfeq_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfeq_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfeq_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfeq_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfeq_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfeq_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfeq_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfeq_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfeq_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfeq.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfeq_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfeq_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfge.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfge.c
new file mode 100644
index 0000000000000..82586da09b325
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfge.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfge_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfge_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfge_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfge_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfge_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfge_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfge_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfge_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfge_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfge_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfge_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfge.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfge_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfge_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfgt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfgt.c
new file mode 100644
index 0000000000000..75ccbbc1e8a1e
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfgt.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfgt_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfgt_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfgt_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfgt_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfgt_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfgt_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfgt_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfgt_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfgt_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfgt_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfgt_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfgt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfgt_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfgt_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfle.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfle.c
new file mode 100644
index 0000000000000..49ff1c9812d62
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfle.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfle_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfle_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfle_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfle_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfle_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfle_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfle_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfle_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfle_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfle_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfle_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfle.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfle_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfle_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmflt.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmflt.c
new file mode 100644
index 0000000000000..24b3f9c16d943
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmflt.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmflt_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmflt_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmflt_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmflt_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmflt_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmflt_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmflt_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmflt_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmflt_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmflt_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmflt_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmflt.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmflt_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmflt_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfne.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfne.c
new file mode 100644
index 0000000000000..ca3e134910dfb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-autogenerated/zvfbfa/policy/overloaded/vmfne.c
@@ -0,0 +1,129 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// REQUIRES: riscv-registered-target
+// RUN: %clang_cc1 -triple riscv64 -target-feature +v \
+// RUN:   -target-feature +experimental-zvfbfa -disable-O0-optnone \
+// RUN:   -emit-llvm %s -o - | opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vv_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], <vscale x 1 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.nxv1bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], <vscale x 1 x bfloat> [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vv_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, vbfloat16mf4_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 1 x i1> @test_vmfne_vf_bf16mf4_b64_mu(
+// CHECK-RV64-SAME: <vscale x 1 x i1> [[MASK:%.*]], <vscale x 1 x i1> [[MASKEDOFF:%.*]], <vscale x 1 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16.i64(<vscale x 1 x i1> [[MASKEDOFF]], <vscale x 1 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 1 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 1 x i1> [[TMP0]]
+//
+vbool64_t test_vmfne_vf_bf16mf4_b64_mu(vbool64_t mask, vbool64_t maskedoff, vbfloat16mf4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vv_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], <vscale x 2 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.nxv2bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], <vscale x 2 x bfloat> [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vv_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, vbfloat16mf2_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i1> @test_vmfne_vf_bf16mf2_b32_mu(
+// CHECK-RV64-SAME: <vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i1> [[MASKEDOFF:%.*]], <vscale x 2 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16.i64(<vscale x 2 x i1> [[MASKEDOFF]], <vscale x 2 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i1> [[TMP0]]
+//
+vbool32_t test_vmfne_vf_bf16mf2_b32_mu(vbool32_t mask, vbool32_t maskedoff, vbfloat16mf2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vv_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], <vscale x 4 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.nxv4bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], <vscale x 4 x bfloat> [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vv_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, vbfloat16m1_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i1> @test_vmfne_vf_bf16m1_b16_mu(
+// CHECK-RV64-SAME: <vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i1> [[MASKEDOFF:%.*]], <vscale x 4 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16.i64(<vscale x 4 x i1> [[MASKEDOFF]], <vscale x 4 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i1> [[TMP0]]
+//
+vbool16_t test_vmfne_vf_bf16m1_b16_mu(vbool16_t mask, vbool16_t maskedoff, vbfloat16m1_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vv_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], <vscale x 8 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.nxv8bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], <vscale x 8 x bfloat> [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vv_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, vbfloat16m2_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i1> @test_vmfne_vf_bf16m2_b8_mu(
+// CHECK-RV64-SAME: <vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i1> [[MASKEDOFF:%.*]], <vscale x 8 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16.i64(<vscale x 8 x i1> [[MASKEDOFF]], <vscale x 8 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i1> [[TMP0]]
+//
+vbool8_t test_vmfne_vf_bf16m2_b8_mu(vbool8_t mask, vbool8_t maskedoff, vbfloat16m2_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vv_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], <vscale x 16 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.nxv16bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], <vscale x 16 x bfloat> [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vv_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, vbfloat16m4_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i1> @test_vmfne_vf_bf16m4_b4_mu(
+// CHECK-RV64-SAME: <vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i1> [[MASKEDOFF:%.*]], <vscale x 16 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16.i64(<vscale x 16 x i1> [[MASKEDOFF]], <vscale x 16 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i1> [[TMP0]]
+//
+vbool4_t test_vmfne_vf_bf16m4_b4_mu(vbool4_t mask, vbool4_t maskedoff, vbfloat16m4_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vv_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], <vscale x 32 x bfloat> [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.nxv32bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], <vscale x 32 x bfloat> [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vv_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, vbfloat16m8_t op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i1> @test_vmfne_vf_bf16m8_b2_mu(
+// CHECK-RV64-SAME: <vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i1> [[MASKEDOFF:%.*]], <vscale x 32 x bfloat> [[OP1:%.*]], bfloat noundef [[OP2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0]] {
+// CHECK-RV64-NEXT:  [[ENTRY:.*:]]
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i1> @llvm.riscv.vmfne.mask.nxv32bf16.bf16.i64(<vscale x 32 x i1> [[MASKEDOFF]], <vscale x 32 x bfloat> [[OP1]], bfloat [[OP2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i1> [[TMP0]]
+//
+vbool2_t test_vmfne_vf_bf16m8_b2_mu(vbool2_t mask, vbool2_t maskedoff, vbfloat16m8_t op1, __bf16 op2, size_t vl) {
+  return __riscv_vmfne_mu(mask, maskedoff, op1, op2, vl);
+}
+
diff --git a/llvm/lib/Target/RISCV/MCTargetDesc/RISCVBaseInfo.h b/llvm/lib/Target/RISCV/MCTargetDesc/RISCVBaseInfo.h
index 70b7c430c410e..649fdb0791664 100644
--- a/llvm/lib/Target/RISCV/MCTargetDesc/RISCVBaseInfo.h
+++ b/llvm/lib/Target/RISCV/MCTargetDesc/RISCVBaseInfo.h
@@ -142,6 +142,12 @@ enum {
 
   ReadsPastVLShift = DestEEWShift + 2,
   ReadsPastVLMask = 1ULL << ReadsPastVLShift,
+
+  // 0 -> Don't care about altfmt bit in VTYPE.
+  // 1 -> Is not altfmt.
+  // 2 -> Is altfmt(BF16).
+  AltFmtTypeShift = ReadsPastVLShift + 1,
+  AltFmtTypeMask = 3ULL << AltFmtTypeShift,
 };
 
 // Helper functions to read TSFlags.
@@ -183,6 +189,11 @@ static inline bool hasRoundModeOp(uint64_t TSFlags) {
   return TSFlags & HasRoundModeOpMask;
 }
 
+enum class AltFmtType { DontCare, IsNotAltFmt, IsAltFmt };
+static inline AltFmtType getAltFmtType(uint64_t TSFlags) {
+  return static_cast<AltFmtType>((TSFlags & AltFmtTypeMask) >> AltFmtTypeShift);
+}
+
 /// \returns true if this instruction uses vxrm
 static inline bool usesVXRM(uint64_t TSFlags) { return TSFlags & UsesVXRMMask; }
 
diff --git a/llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp b/llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp
index 90e1c47a71c89..5a90f469165fb 100644
--- a/llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp
+++ b/llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp
@@ -164,6 +164,7 @@ struct DemandedFields {
   // If this is true, we demand that VTYPE is set to some legal state, i.e. that
   // vill is unset.
   bool VILL = false;
+  bool AltFmt = false;
 
   // Return true if any part of VTYPE was used
   bool usedVTYPE() const {
@@ -183,6 +184,7 @@ struct DemandedFields {
     TailPolicy = true;
     MaskPolicy = true;
     VILL = true;
+    AltFmt = true;
   }
 
   // Mark all VL properties as demanded
@@ -208,6 +210,7 @@ struct DemandedFields {
     TailPolicy |= B.TailPolicy;
     MaskPolicy |= B.MaskPolicy;
     VILL |= B.VILL;
+    AltFmt |= B.AltFmt;
   }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
@@ -324,6 +327,9 @@ static bool areCompatibleVTYPEs(uint64_t CurVType, uint64_t NewVType,
   if (Used.MaskPolicy && RISCVVType::isMaskAgnostic(CurVType) !=
                              RISCVVType::isMaskAgnostic(NewVType))
     return false;
+  if (Used.AltFmt &&
+      RISCVVType::isAltFmt(CurVType) != RISCVVType::isAltFmt(NewVType))
+    return false;
   return true;
 }
 
@@ -475,6 +481,9 @@ DemandedFields getDemanded(const MachineInstr &MI, const RISCVSubtarget *ST) {
     Res.TailPolicy = false;
   }
 
+  Res.AltFmt = RISCVII::getAltFmtType(MI.getDesc().TSFlags) !=
+               RISCVII::AltFmtType::DontCare;
+
   return Res;
 }
 
@@ -506,6 +515,7 @@ class VSETVLIInfo {
   uint8_t TailAgnostic : 1;
   uint8_t MaskAgnostic : 1;
   uint8_t SEWLMULRatioOnly : 1;
+  uint8_t AltFmt : 1;
 
 public:
   VSETVLIInfo()
@@ -582,6 +592,7 @@ class VSETVLIInfo {
   RISCVVType::VLMUL getVLMUL() const { return VLMul; }
   bool getTailAgnostic() const { return TailAgnostic; }
   bool getMaskAgnostic() const { return MaskAgnostic; }
+  bool getAltFmt() const { return AltFmt; }
 
   bool hasNonZeroAVL(const LiveIntervals *LIS) const {
     if (hasAVLImm())
@@ -643,14 +654,16 @@ class VSETVLIInfo {
     SEW = RISCVVType::getSEW(VType);
     TailAgnostic = RISCVVType::isTailAgnostic(VType);
     MaskAgnostic = RISCVVType::isMaskAgnostic(VType);
+    AltFmt = RISCVVType::isAltFmt(VType);
   }
-  void setVTYPE(RISCVVType::VLMUL L, unsigned S, bool TA, bool MA) {
+  void setVTYPE(RISCVVType::VLMUL L, unsigned S, bool TA, bool MA, bool AF) {
     assert(isValid() && !isUnknown() &&
            "Can't set VTYPE for uninitialized or unknown");
     VLMul = L;
     SEW = S;
     TailAgnostic = TA;
     MaskAgnostic = MA;
+    AltFmt = AF;
   }
 
   void setVLMul(RISCVVType::VLMUL VLMul) { this->VLMul = VLMul; }
@@ -658,7 +671,8 @@ class VSETVLIInfo {
   unsigned encodeVTYPE() const {
     assert(isValid() && !isUnknown() && !SEWLMULRatioOnly &&
            "Can't encode VTYPE for uninitialized or unknown");
-    return RISCVVType::encodeVTYPE(VLMul, SEW, TailAgnostic, MaskAgnostic);
+    return RISCVVType::encodeVTYPE(VLMul, SEW, TailAgnostic, MaskAgnostic,
+                                   AltFmt);
   }
 
   bool hasSEWLMULRatioOnly() const { return SEWLMULRatioOnly; }
@@ -670,9 +684,9 @@ class VSETVLIInfo {
            "Can't compare VTYPE in unknown state");
     assert(!SEWLMULRatioOnly && !Other.SEWLMULRatioOnly &&
            "Can't compare when only LMUL/SEW ratio is valid.");
-    return std::tie(VLMul, SEW, TailAgnostic, MaskAgnostic) ==
+    return std::tie(VLMul, SEW, TailAgnostic, MaskAgnostic, AltFmt) ==
            std::tie(Other.VLMul, Other.SEW, Other.TailAgnostic,
-                    Other.MaskAgnostic);
+                    Other.MaskAgnostic, Other.AltFmt);
   }
 
   unsigned getSEWLMULRatio() const {
@@ -1001,6 +1015,8 @@ RISCVInsertVSETVLI::computeInfoForInstr(const MachineInstr &MI) const {
 
   RISCVVType::VLMUL VLMul = RISCVII::getLMul(TSFlags);
 
+  bool AltFmt =
+      RISCVII::getAltFmtType(TSFlags) == RISCVII::AltFmtType::IsAltFmt;
   unsigned Log2SEW = MI.getOperand(getSEWOpNum(MI)).getImm();
   // A Log2SEW of 0 is an operation on mask registers only.
   unsigned SEW = Log2SEW ? 1 << Log2SEW : 8;
@@ -1041,7 +1057,7 @@ RISCVInsertVSETVLI::computeInfoForInstr(const MachineInstr &MI) const {
     assert(SEW == EEW && "Initial SEW doesn't match expected EEW");
   }
 #endif
-  InstrInfo.setVTYPE(VLMul, SEW, TailAgnostic, MaskAgnostic);
+  InstrInfo.setVTYPE(VLMul, SEW, TailAgnostic, MaskAgnostic, AltFmt);
 
   forwardVSETVLIAVL(InstrInfo);
 
@@ -1194,7 +1210,8 @@ void RISCVInsertVSETVLI::transferBefore(VSETVLIInfo &Info,
     // be coalesced into another vsetvli since we won't demand any fields.
     VSETVLIInfo NewInfo; // Need a new VSETVLIInfo to clear SEWLMULRatioOnly
     NewInfo.setAVLImm(1);
-    NewInfo.setVTYPE(RISCVVType::LMUL_1, /*sew*/ 8, /*ta*/ true, /*ma*/ true);
+    NewInfo.setVTYPE(RISCVVType::LMUL_1, /*sew*/ 8, /*ta*/ true, /*ma*/ true,
+                     /*AF*/ false);
     Info = NewInfo;
     return;
   }
@@ -1236,7 +1253,8 @@ void RISCVInsertVSETVLI::transferBefore(VSETVLIInfo &Info,
       (Demanded.TailPolicy ? IncomingInfo : Info).getTailAgnostic() ||
           IncomingInfo.getTailAgnostic(),
       (Demanded.MaskPolicy ? IncomingInfo : Info).getMaskAgnostic() ||
-          IncomingInfo.getMaskAgnostic());
+          IncomingInfo.getMaskAgnostic(),
+      Demanded.AltFmt ? IncomingInfo.getAltFmt() : 0);
 
   // If we only knew the sew/lmul ratio previously, replace the VTYPE but keep
   // the AVL.
diff --git a/llvm/lib/Target/RISCV/RISCVInstrFormats.td b/llvm/lib/Target/RISCV/RISCVInstrFormats.td
index 2afd77a96373b..7b262f2950ba5 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrFormats.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrFormats.td
@@ -267,6 +267,12 @@ class RVInstCommon<dag outs, dag ins, string opcodestr, string argstr,
   // operands' VLs.
   bit ReadsPastVL = 0;
   let TSFlags{26} = ReadsPastVL;
+
+  // 0 -> Don't care about altfmt bit in VTYPE.
+  // 1 -> Is not altfmt.
+  // 2 -> Is altfmt(BF16).
+  bits<2> AltFmtType = 0;
+  let TSFlags{28-27} = AltFmtType;
 }
 
 class RVInst<dag outs, dag ins, string opcodestr, string argstr,
diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
index d998316e4725c..9950f7f176f11 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td
@@ -128,6 +128,9 @@ defvar TAIL_AGNOSTIC = 1;
 defvar TU_MU = 0;
 defvar TA_MU = 1;
 defvar TA_MA = 3;
+defvar DONT_CARE_ALTFMT = 0;
+defvar IS_NOT_ALTFMT = 1;
+defvar IS_ALTFMT = 2;
 
 //===----------------------------------------------------------------------===//
 // Utilities.
@@ -159,7 +162,8 @@ class PseudoToVInst<string PseudoInst> {
                         ["_M4", ""],
                         ["_M8", ""],
                         ["_SE", ""],
-                        ["_RM", ""]
+                        ["_RM", ""],
+                        ["_ALT", ""]
                        ];
   string VInst = !foldl(PseudoInst, AffixSubsts, Acc, AffixSubst,
                         !subst(AffixSubst[0], AffixSubst[1], Acc));
@@ -6403,7 +6407,7 @@ let Defs = [VXSAT] in {
 // 13. Vector Floating-Point Instructions
 //===----------------------------------------------------------------------===//
 
-let Predicates = [HasVInstructionsAnyF] in {
+let Predicates = [HasVInstructionsAnyF], AltFmtType = IS_NOT_ALTFMT in {
 //===----------------------------------------------------------------------===//
 // 13.2. Vector Single-Width Floating-Point Add/Subtract Instructions
 //===----------------------------------------------------------------------===//
@@ -6600,7 +6604,7 @@ defm PseudoVWREDSUM    : VPseudoVWRED_VS;
 }
 } // Predicates = [HasVInstructions]
 
-let Predicates = [HasVInstructionsAnyF] in {
+let Predicates = [HasVInstructionsAnyF], AltFmtType = IS_NOT_ALTFMT in {
 //===----------------------------------------------------------------------===//
 // 14.3. Vector Single-Width Floating-Point Reduction Instructions
 //===----------------------------------------------------------------------===//
@@ -6710,7 +6714,7 @@ let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in {
 // 16.2. Floating-Point Scalar Move Instructions
 //===----------------------------------------------------------------------===//
 
-let Predicates = [HasVInstructionsAnyF] in {
+let Predicates = [HasVInstructionsAnyF], AltFmtType = IS_NOT_ALTFMT in {
 let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in {
   foreach f = FPList in {
     let HasSEWOp = 1, BaseInstr = VFMV_F_S in
@@ -6738,7 +6742,7 @@ let Predicates = [HasVInstructions] in {
   defm PseudoVSLIDE1DOWN : VPseudoVSLD1_VX;
 } // Predicates = [HasVInstructions]
 
-let Predicates = [HasVInstructionsAnyF] in {
+let Predicates = [HasVInstructionsAnyF], AltFmtType = IS_NOT_ALTFMT in {
   defm PseudoVFSLIDE1UP  : VPseudoVSLD1_VF<"@earlyclobber $rd">;
   defm PseudoVFSLIDE1DOWN : VPseudoVSLD1_VF;
 } // Predicates = [HasVInstructionsAnyF]
diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoZvfbf.td b/llvm/lib/Target/RISCV/RISCVInstrInfoZvfbf.td
index 0be9eab6870ec..a346b763b7941 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfoZvfbf.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfoZvfbf.td
@@ -44,6 +44,336 @@ let Predicates = [HasStdExtZvfbfminOrZvfofp8min] in {
 let mayRaiseFPException = true, Predicates = [HasStdExtZvfbfwma] in
   defm PseudoVFWMACCBF16 : VPseudoVWMAC_VV_VF_BF_RM;
 
+defset list<VTypeInfoToWide> AllWidenableIntToBF16Vectors = {
+  def : VTypeInfoToWide<VI8MF8, VBF16MF4>;
+  def : VTypeInfoToWide<VI8MF4, VBF16MF2>;
+  def : VTypeInfoToWide<VI8MF2, VBF16M1>;
+  def : VTypeInfoToWide<VI8M1, VBF16M2>;
+  def : VTypeInfoToWide<VI8M2, VBF16M4>;
+  def : VTypeInfoToWide<VI8M4, VBF16M8>;
+}
+
+multiclass VPseudoVALU_VV_VF_RM_BF16 {
+  foreach m = MxListF in {
+    defm "" : VPseudoBinaryFV_VV_RM<m, 16/*sew*/>,
+              SchedBinary<"WriteVFALUV", "ReadVFALUV", "ReadVFALUV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoBinaryV_VF_RM<m, f, f.SEW>,
+              SchedBinary<"WriteVFALUF", "ReadVFALUV", "ReadVFALUF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVALU_VF_RM_BF16 {
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoBinaryV_VF_RM<m, f, f.SEW>,
+              SchedBinary<"WriteVFALUF", "ReadVFALUV", "ReadVFALUF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVFWALU_VV_VF_RM_BF16 {
+  foreach m = MxListFW in {
+    defm "" : VPseudoBinaryW_VV_RM<m, sew=16>,
+              SchedBinary<"WriteVFWALUV", "ReadVFWALUV", "ReadVFWALUV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxListFW in {
+    defm "" : VPseudoBinaryW_VF_RM<m, f, sew=f.SEW>,
+              SchedBinary<"WriteVFWALUF", "ReadVFWALUV", "ReadVFWALUF", m.MX,
+                        f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVFWALU_WV_WF_RM_BF16 {
+  foreach m = MxListFW in {
+    defm "" : VPseudoBinaryW_WV_RM<m, sew=16>,
+              SchedBinary<"WriteVFWALUV", "ReadVFWALUV", "ReadVFWALUV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+  defvar f = SCALAR_F16;
+  foreach m = f.MxListFW in {
+    defm "" : VPseudoBinaryW_WF_RM<m, f, sew=f.SEW>,
+              SchedBinary<"WriteVFWALUF", "ReadVFWALUV", "ReadVFWALUF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVFMUL_VV_VF_RM_BF16 {
+  foreach m = MxListF in {
+    defm "" : VPseudoBinaryFV_VV_RM<m, 16/*sew*/>,
+              SchedBinary<"WriteVFMulV", "ReadVFMulV", "ReadVFMulV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoBinaryV_VF_RM<m, f, f.SEW>,
+              SchedBinary<"WriteVFMulF", "ReadVFMulV", "ReadVFMulF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVWMUL_VV_VF_RM_BF16 {
+  foreach m = MxListFW in {
+    defm "" : VPseudoBinaryW_VV_RM<m, sew=16>,
+              SchedBinary<"WriteVFWMulV", "ReadVFWMulV", "ReadVFWMulV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxListFW in {
+    defm "" : VPseudoBinaryW_VF_RM<m, f, sew=f.SEW>,
+              SchedBinary<"WriteVFWMulF", "ReadVFWMulV", "ReadVFWMulF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVMAC_VV_VF_AAXA_RM_BF16 {
+  foreach m = MxListF in {
+    defm "" : VPseudoTernaryV_VV_AAXA_RM<m, 16/*sew*/>,
+              SchedTernary<"WriteVFMulAddV", "ReadVFMulAddV", "ReadVFMulAddV", 
+                           "ReadVFMulAddV", m.MX, 16/*sew*/>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoTernaryV_VF_AAXA_RM<m, f, f.SEW>,
+              SchedTernary<"WriteVFMulAddF", "ReadVFMulAddV", "ReadVFMulAddF", 
+                           "ReadVFMulAddV", m.MX, f.SEW>;
+  }
+}
+
+multiclass VPseudoVWMAC_VV_VF_RM_BF16 {
+  foreach m = MxListFW in {
+    defm "" : VPseudoTernaryW_VV_RM<m, sew=16>,
+              SchedTernary<"WriteVFWMulAddV", "ReadVFWMulAddV",
+                           "ReadVFWMulAddV", "ReadVFWMulAddV", m.MX, 16/*sew*/>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxListFW in {
+    defm "" : VPseudoTernaryW_VF_RM<m, f, sew=f.SEW>,
+              SchedTernary<"WriteVFWMulAddF", "ReadVFWMulAddV",
+                           "ReadVFWMulAddF", "ReadVFWMulAddV", m.MX, f.SEW>;
+  }
+}
+
+multiclass VPseudoVRCP_V_BF16 {
+  foreach m = MxListF in {
+    defvar mx = m.MX;
+    let VLMul = m.value in {
+      def "_V_" # mx # "_E16"
+          : VPseudoUnaryNoMask<m.vrclass, m.vrclass>,
+            SchedUnary<"WriteVFRecpV", "ReadVFRecpV", mx, 16/*sew*/,
+                       forcePassthruRead=true>;
+      def "_V_" # mx # "_E16_MASK"
+          : VPseudoUnaryMask<m.vrclass, m.vrclass>,
+            RISCVMaskedPseudo<MaskIdx = 2>,
+            SchedUnary<"WriteVFRecpV", "ReadVFRecpV", mx, 16/*sew*/,
+                       forcePassthruRead=true>;
+    }
+  }
+}
+
+multiclass VPseudoVRCP_V_RM_BF16 {
+  foreach m = MxListF in {
+    defvar mx = m.MX;
+    let VLMul = m.value in {
+      def "_V_" # mx # "_E16"
+          : VPseudoUnaryNoMaskRoundingMode<m.vrclass, m.vrclass>,
+            SchedUnary<"WriteVFRecpV", "ReadVFRecpV", mx, 16/*sew*/,
+                       forcePassthruRead=true>;
+      def "_V_" # mx # "_E16_MASK"
+          : VPseudoUnaryMaskRoundingMode<m.vrclass, m.vrclass>,
+            RISCVMaskedPseudo<MaskIdx = 2>,
+            SchedUnary<"WriteVFRecpV", "ReadVFRecpV", mx, 16/*sew*/,
+                       forcePassthruRead=true>;
+    }
+  }
+}
+
+multiclass VPseudoVMAX_VV_VF_BF16 {
+  foreach m = MxListF in {
+    defm "" : VPseudoBinaryV_VV<m, sew=16>,
+              SchedBinary<"WriteVFMinMaxV", "ReadVFMinMaxV", "ReadVFMinMaxV", 
+                          m.MX, 16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoBinaryV_VF<m, f, f.SEW>,
+              SchedBinary<"WriteVFMinMaxF", "ReadVFMinMaxV", "ReadVFMinMaxF", 
+                          m.MX, f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVSGNJ_VV_VF_BF16 {
+  foreach m = MxListF in {
+    defm "" : VPseudoBinaryV_VV<m, sew=16>,
+              SchedBinary<"WriteVFSgnjV", "ReadVFSgnjV", "ReadVFSgnjV", m.MX,
+                          16/*sew*/, forcePassthruRead=true>;
+  }
+
+  defvar f = SCALAR_F16;
+  foreach m = f.MxList in {
+    defm "" : VPseudoBinaryV_VF<m, f, f.SEW>,
+              SchedBinary<"WriteVFSgnjF", "ReadVFSgnjV", "ReadVFSgnjF", m.MX,
+                          f.SEW, forcePassthruRead=true>;
+  }
+}
+
+multiclass VPseudoVWCVTF_V_BF16 {
+  defvar constraint = "@earlyclobber $rd";
+  foreach m = MxListW in
+    defm _V : VPseudoConversion<m.wvrclass, m.vrclass, m, constraint, sew=8,
+                                TargetConstraintType=3>,
+              SchedUnary<"WriteVFWCvtIToFV", "ReadVFWCvtIToFV", m.MX, 8/*sew*/,
+                         forcePassthruRead=true>;
+}
+
+multiclass VPseudoVWCVTD_V_BF16 {
+  defvar constraint = "@earlyclobber $rd";
+  foreach m = MxListFW in
+    defm _V : VPseudoConversion<m.wvrclass, m.vrclass, m, constraint, sew=16,
+                                TargetConstraintType=3>,
+              SchedUnary<"WriteVFWCvtFToFV", "ReadVFWCvtFToFV", m.MX, 16/*sew*/,
+                         forcePassthruRead=true>;
+}
+
+multiclass VPseudoVNCVTD_W_BF16 {
+  defvar constraint = "@earlyclobber $rd";
+  foreach m = MxListFW in
+    defm _W : VPseudoConversion<m.vrclass, m.wvrclass, m, constraint, sew=16,
+                                TargetConstraintType=2>,
+              SchedUnary<"WriteVFNCvtFToFV", "ReadVFNCvtFToFV", m.MX, 16/*sew*/,
+                         forcePassthruRead=true>;
+}
+
+multiclass VPseudoVNCVTD_W_RM_BF16 {
+  defvar constraint = "@earlyclobber $rd";
+  foreach m = MxListFW in
+    defm _W : VPseudoConversionRoundingMode<m.vrclass, m.wvrclass, m,
+                                            constraint, sew=16,
+                                            TargetConstraintType=2>,
+              SchedUnary<"WriteVFNCvtFToFV", "ReadVFNCvtFToFV", m.MX, 16/*sew*/,
+                         forcePassthruRead=true>;
+}
+
+let Predicates = [HasStdExtZvfbfa], AltFmtType = IS_ALTFMT in {
+let mayRaiseFPException = true in {
+defm PseudoVFADD_ALT : VPseudoVALU_VV_VF_RM_BF16;
+defm PseudoVFSUB_ALT  : VPseudoVALU_VV_VF_RM_BF16;
+defm PseudoVFRSUB_ALT : VPseudoVALU_VF_RM_BF16;
+}
+
+let mayRaiseFPException = true in {
+defm PseudoVFWADD_ALT : VPseudoVFWALU_VV_VF_RM_BF16;
+defm PseudoVFWSUB_ALT : VPseudoVFWALU_VV_VF_RM_BF16;
+defm PseudoVFWADD_ALT : VPseudoVFWALU_WV_WF_RM_BF16;
+defm PseudoVFWSUB_ALT : VPseudoVFWALU_WV_WF_RM_BF16;
+}
+
+let mayRaiseFPException = true in
+defm PseudoVFMUL_ALT : VPseudoVFMUL_VV_VF_RM_BF16;
+
+let mayRaiseFPException = true in
+defm PseudoVFWMUL_ALT : VPseudoVWMUL_VV_VF_RM_BF16;
+
+let mayRaiseFPException = true in {
+defm PseudoVFMACC_ALT  : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFNMACC_ALT : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFMSAC_ALT  : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFNMSAC_ALT : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFMADD_ALT  : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFNMADD_ALT : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFMSUB_ALT  : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+defm PseudoVFNMSUB_ALT : VPseudoVMAC_VV_VF_AAXA_RM_BF16;
+}
+
+let mayRaiseFPException = true in {
+defm PseudoVFWMACC_ALT  : VPseudoVWMAC_VV_VF_RM_BF16;
+defm PseudoVFWNMACC_ALT : VPseudoVWMAC_VV_VF_RM_BF16;
+defm PseudoVFWMSAC_ALT  : VPseudoVWMAC_VV_VF_RM_BF16;
+defm PseudoVFWNMSAC_ALT : VPseudoVWMAC_VV_VF_RM_BF16;
+}
+
+let mayRaiseFPException = true in
+defm PseudoVFRSQRT7_ALT : VPseudoVRCP_V_BF16;
+
+let mayRaiseFPException = true in
+defm PseudoVFREC7_ALT : VPseudoVRCP_V_RM_BF16;
+
+let mayRaiseFPException = true in {
+defm PseudoVFMIN_ALT : VPseudoVMAX_VV_VF_BF16;
+defm PseudoVFMAX_ALT : VPseudoVMAX_VV_VF_BF16;
+}
+
+defm PseudoVFSGNJ_ALT  : VPseudoVSGNJ_VV_VF_BF16;
+defm PseudoVFSGNJN_ALT : VPseudoVSGNJ_VV_VF_BF16;
+defm PseudoVFSGNJX_ALT : VPseudoVSGNJ_VV_VF_BF16;
+
+let mayRaiseFPException = true in {
+defm PseudoVMFEQ_ALT : VPseudoVCMPM_VV_VF;
+defm PseudoVMFNE_ALT : VPseudoVCMPM_VV_VF;
+defm PseudoVMFLT_ALT : VPseudoVCMPM_VV_VF;
+defm PseudoVMFLE_ALT : VPseudoVCMPM_VV_VF;
+defm PseudoVMFGT_ALT : VPseudoVCMPM_VF;
+defm PseudoVMFGE_ALT : VPseudoVCMPM_VF;
+}
+
+defm PseudoVFCLASS_ALT : VPseudoVCLS_V;
+
+defm PseudoVFMERGE_ALT : VPseudoVMRG_FM;
+
+defm PseudoVFMV_V_ALT : VPseudoVMV_F;
+
+let mayRaiseFPException = true in {
+defm PseudoVFWCVT_F_XU_ALT : VPseudoVWCVTF_V_BF16;
+defm PseudoVFWCVT_F_X_ALT  : VPseudoVWCVTF_V_BF16;
+
+defm PseudoVFWCVT_F_F_ALT  : VPseudoVWCVTD_V_BF16;
+} // mayRaiseFPException = true
+
+let mayRaiseFPException = true in {
+let hasSideEffects = 0, hasPostISelHook = 1 in {
+defm PseudoVFNCVT_XU_F_ALT : VPseudoVNCVTI_W_RM;
+defm PseudoVFNCVT_X_F_ALT  : VPseudoVNCVTI_W_RM;
+}
+
+defm PseudoVFNCVT_RTZ_XU_F_ALT : VPseudoVNCVTI_W;
+defm PseudoVFNCVT_RTZ_X_F_ALT  : VPseudoVNCVTI_W;
+
+defm PseudoVFNCVT_F_F_ALT  : VPseudoVNCVTD_W_RM_BF16;
+
+defm PseudoVFNCVT_ROD_F_F_ALT : VPseudoVNCVTD_W_BF16;
+} // mayRaiseFPException = true
+
+let mayLoad = 0, mayStore = 0, hasSideEffects = 0 in {
+  defvar f = SCALAR_F16;
+  let HasSEWOp = 1, BaseInstr = VFMV_F_S in
+  def "PseudoVFMV_" # f.FX # "_S_ALT" :
+    RISCVVPseudo<(outs f.fprclass:$rd), (ins VR:$rs2, sew:$sew)>,
+    Sched<[WriteVMovFS, ReadVMovFS]>;
+  let HasVLOp = 1, HasSEWOp = 1, BaseInstr = VFMV_S_F, isReMaterializable = 1,
+      Constraints = "$rd = $passthru" in
+  def "PseudoVFMV_S_" # f.FX # "_ALT" :
+    RISCVVPseudo<(outs VR:$rd),
+                 (ins VR:$passthru, f.fprclass:$rs1, AVL:$vl, sew:$sew)>,
+    Sched<[WriteVMovSF, ReadVMovSF_V, ReadVMovSF_F]>;
+}
+
+defm PseudoVFSLIDE1UP_ALT   : VPseudoVSLD1_VF<"@earlyclobber $rd">;
+defm PseudoVFSLIDE1DOWN_ALT : VPseudoVSLD1_VF;
+} // Predicates = [HasStdExtZvfbfa], AltFmtType = IS_ALTFMT
+
 //===----------------------------------------------------------------------===//
 // Patterns
 //===----------------------------------------------------------------------===//
@@ -87,6 +417,130 @@ let Predicates = [HasStdExtZvfbfminOrZvfofp8min] in {
                   FRM_DYN,
                   fvti.AVL, fvti.Log2SEW, TA_MA)>;
   }
+
+  defm : VPatUnaryV_V_AnyMask<"int_riscv_vcompress", "PseudoVCOMPRESS", AllBF16Vectors>;
+  defm : VPatBinaryV_VV_VX_VI_INT<"int_riscv_vrgather", "PseudoVRGATHER",
+                                  AllBF16Vectors, uimm5>;
+  defm : VPatBinaryV_VV_INT_EEW<"int_riscv_vrgatherei16_vv", "PseudoVRGATHEREI16",
+                                eew=16, vtilist=AllBF16Vectors>;
+  defm : VPatTernaryV_VX_VI<"int_riscv_vslideup", "PseudoVSLIDEUP", AllBF16Vectors, uimm5>;
+  defm : VPatTernaryV_VX_VI<"int_riscv_vslidedown", "PseudoVSLIDEDOWN", AllBF16Vectors, uimm5>;
+
+  foreach fvti = AllBF16Vectors in {
+    defm : VPatBinaryCarryInTAIL<"int_riscv_vmerge", "PseudoVMERGE", "VVM",
+                                 fvti.Vector,
+                                 fvti.Vector, fvti.Vector, fvti.Mask,
+                                 fvti.Log2SEW, fvti.LMul, fvti.RegClass,
+                                 fvti.RegClass, fvti.RegClass>;
+    defm : VPatBinaryCarryInTAIL<"int_riscv_vfmerge", "PseudoVFMERGE",
+                                 "V"#fvti.ScalarSuffix#"M",
+                                 fvti.Vector,
+                                 fvti.Vector, fvti.Scalar, fvti.Mask,
+                                 fvti.Log2SEW, fvti.LMul, fvti.RegClass,
+                                 fvti.RegClass, fvti.ScalarRegClass>;
+    defvar instr = !cast<Instruction>("PseudoVMERGE_VIM_"#fvti.LMul.MX);
+    def : Pat<(fvti.Vector (int_riscv_vfmerge (fvti.Vector fvti.RegClass:$passthru),
+                                              (fvti.Vector fvti.RegClass:$rs2),
+                                              (fvti.Scalar (fpimm0)),
+                                              (fvti.Mask VMV0:$vm), VLOpFrag)),
+              (instr fvti.RegClass:$passthru, fvti.RegClass:$rs2, 0,
+                     (fvti.Mask VMV0:$vm), GPR:$vl, fvti.Log2SEW)>;
+
+    defvar ivti = GetIntVTypeInfo<fvti>.Vti;
+    def : Pat<(fvti.Vector (vselect (fvti.Mask VMV0:$vm), fvti.RegClass:$rs1,
+                                                          fvti.RegClass:$rs2)),
+              (!cast<Instruction>("PseudoVMERGE_VVM_"#fvti.LMul.MX)
+                   (fvti.Vector (IMPLICIT_DEF)),
+                   fvti.RegClass:$rs2, fvti.RegClass:$rs1, (fvti.Mask VMV0:$vm),
+                   fvti.AVL, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (vselect (fvti.Mask VMV0:$vm),
+                                    (SplatFPOp (SelectScalarFPAsInt (XLenVT GPR:$imm))),
+                                    fvti.RegClass:$rs2)),
+              (!cast<Instruction>("PseudoVMERGE_VXM_"#fvti.LMul.MX)
+                   (fvti.Vector (IMPLICIT_DEF)),
+                   fvti.RegClass:$rs2, GPR:$imm, (fvti.Mask VMV0:$vm), fvti.AVL, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (vselect (fvti.Mask VMV0:$vm),
+                                    (SplatFPOp (fvti.Scalar fpimm0)),
+                                    fvti.RegClass:$rs2)),
+              (!cast<Instruction>("PseudoVMERGE_VIM_"#fvti.LMul.MX)
+                   (fvti.Vector (IMPLICIT_DEF)),
+                   fvti.RegClass:$rs2, 0, (fvti.Mask VMV0:$vm), fvti.AVL, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (vselect (fvti.Mask VMV0:$vm),
+                                    (SplatFPOp fvti.ScalarRegClass:$rs1),
+                                    fvti.RegClass:$rs2)),
+              (!cast<Instruction>("PseudoVFMERGE_V"#fvti.ScalarSuffix#"M_"#fvti.LMul.MX)
+                   (fvti.Vector (IMPLICIT_DEF)),
+                   fvti.RegClass:$rs2,
+                   (fvti.Scalar fvti.ScalarRegClass:$rs1),
+                   (fvti.Mask VMV0:$vm), fvti.AVL, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (riscv_vmerge_vl (fvti.Mask VMV0:$vm),
+                                            fvti.RegClass:$rs1,
+                                            fvti.RegClass:$rs2,
+                                            fvti.RegClass:$passthru,
+                                            VLOpFrag)),
+              (!cast<Instruction>("PseudoVMERGE_VVM_"#fvti.LMul.MX)
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, fvti.RegClass:$rs1, (fvti.Mask VMV0:$vm),
+                   GPR:$vl, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (riscv_vmerge_vl (fvti.Mask VMV0:$vm),
+                                            (SplatFPOp (SelectScalarFPAsInt (XLenVT GPR:$imm))),
+                                            fvti.RegClass:$rs2,
+                                            fvti.RegClass:$passthru,
+                                            VLOpFrag)),
+              (!cast<Instruction>("PseudoVMERGE_VXM_"#fvti.LMul.MX)
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, GPR:$imm, (fvti.Mask VMV0:$vm),
+                   GPR:$vl, fvti.Log2SEW)>;
+
+
+    def : Pat<(fvti.Vector (riscv_vmerge_vl (fvti.Mask VMV0:$vm),
+                                            (SplatFPOp (fvti.Scalar fpimm0)),
+                                            fvti.RegClass:$rs2,
+                                            fvti.RegClass:$passthru,
+                                            VLOpFrag)),
+              (!cast<Instruction>("PseudoVMERGE_VIM_"#fvti.LMul.MX)
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, 0, (fvti.Mask VMV0:$vm),
+                   GPR:$vl, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector (riscv_vmerge_vl (fvti.Mask VMV0:$vm),
+                                            (SplatFPOp fvti.ScalarRegClass:$rs1),
+                                            fvti.RegClass:$rs2,
+                                            fvti.RegClass:$passthru,
+                                            VLOpFrag)),
+              (!cast<Instruction>("PseudoVFMERGE_V"#fvti.ScalarSuffix#"M_"#fvti.LMul.MX)
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2,
+                   (fvti.Scalar fvti.ScalarRegClass:$rs1),
+                   (fvti.Mask VMV0:$vm), GPR:$vl, fvti.Log2SEW)>;
+
+    def : Pat<(fvti.Vector
+               (riscv_vrgather_vv_vl fvti.RegClass:$rs2,
+                                     (ivti.Vector fvti.RegClass:$rs1),
+                                     fvti.RegClass:$passthru,
+                                     (fvti.Mask VMV0:$vm),
+                                     VLOpFrag)),
+              (!cast<Instruction>("PseudoVRGATHER_VV_"# fvti.LMul.MX#"_E"# fvti.SEW#"_MASK")
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, fvti.RegClass:$rs1,
+                   (fvti.Mask VMV0:$vm), GPR:$vl, fvti.Log2SEW, TAIL_AGNOSTIC)>;
+    def : Pat<(fvti.Vector (riscv_vrgather_vx_vl fvti.RegClass:$rs2, GPR:$rs1,
+                                                fvti.RegClass:$passthru,
+                                                (fvti.Mask VMV0:$vm),
+                                                VLOpFrag)),
+              (!cast<Instruction>("PseudoVRGATHER_VX_"# fvti.LMul.MX#"_MASK")
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, GPR:$rs1,
+                   (fvti.Mask VMV0:$vm), GPR:$vl, fvti.Log2SEW, TAIL_AGNOSTIC)>;
+    def : Pat<(fvti.Vector
+               (riscv_vrgather_vx_vl fvti.RegClass:$rs2,
+                                     uimm5:$imm,
+                                     fvti.RegClass:$passthru,
+                                     (fvti.Mask VMV0:$vm),
+                                     VLOpFrag)),
+              (!cast<Instruction>("PseudoVRGATHER_VI_"# fvti.LMul.MX#"_MASK")
+                   fvti.RegClass:$passthru, fvti.RegClass:$rs2, uimm5:$imm,
+                   (fvti.Mask VMV0:$vm), GPR:$vl, fvti.Log2SEW, TAIL_AGNOSTIC)>;
+  }
 }
 
 let Predicates = [HasStdExtZvfbfwma] in {
@@ -97,3 +551,224 @@ let Predicates = [HasStdExtZvfbfwma] in {
   defm : VPatWidenFPMulAccSDNode_VV_VF_RM<"PseudoVFWMACCBF16",
                                           AllWidenableBF16ToFloatVectors>;
 }
+
+multiclass VPatConversionVI_VF_BF16<string intrinsic, string instruction> {
+  foreach fvti = AllBF16Vectors in {
+    defvar ivti = GetIntVTypeInfo<fvti>.Vti;
+    let Predicates = !listconcat(GetVTypePredicates<fvti>.Predicates,
+                                 GetVTypePredicates<ivti>.Predicates) in
+    defm : VPatConversion<intrinsic, instruction, "V",
+                          ivti.Vector, fvti.Vector, ivti.Mask, fvti.Log2SEW,
+                          fvti.LMul, ivti.RegClass, fvti.RegClass>;
+  }
+}
+
+multiclass VPatConversionWF_VI_BF16<string intrinsic, string instruction,
+                                    bit isSEWAware = 0> {
+  foreach vtiToWti = AllWidenableIntToBF16Vectors in {
+    defvar vti = vtiToWti.Vti;
+    defvar fwti = vtiToWti.Wti;
+    let Predicates = !listconcat(GetVTypePredicates<vti>.Predicates,
+                                 GetVTypePredicates<fwti>.Predicates) in
+    defm : VPatConversion<intrinsic, instruction, "V",
+                          fwti.Vector, vti.Vector, fwti.Mask, vti.Log2SEW,
+                          vti.LMul, fwti.RegClass, vti.RegClass, isSEWAware>;
+  }
+}
+
+multiclass VPatConversionWF_VF_BF16<string intrinsic, string instruction,
+                                    bit isSEWAware = 0> {
+  foreach fvtiToFWti = AllWidenableBF16ToFloatVectors in {
+    defvar fvti = fvtiToFWti.Vti;
+    defvar fwti = fvtiToFWti.Wti;
+    let Predicates = !listconcat(GetVTypeMinimalPredicates<fvti>.Predicates,
+                                 GetVTypeMinimalPredicates<fwti>.Predicates) in
+    defm : VPatConversion<intrinsic, instruction, "V",
+                          fwti.Vector, fvti.Vector, fwti.Mask, fvti.Log2SEW,
+                          fvti.LMul, fwti.RegClass, fvti.RegClass, isSEWAware>;
+  }
+}
+
+multiclass VPatConversionVI_WF_BF16<string intrinsic, string instruction> {
+  foreach vtiToWti = AllWidenableIntToBF16Vectors in {
+    defvar vti = vtiToWti.Vti;
+    defvar fwti = vtiToWti.Wti;
+    let Predicates = !listconcat(GetVTypePredicates<vti>.Predicates,
+                                 GetVTypePredicates<fwti>.Predicates) in
+    defm : VPatConversion<intrinsic, instruction, "W",
+                          vti.Vector, fwti.Vector, vti.Mask, vti.Log2SEW,
+                          vti.LMul, vti.RegClass, fwti.RegClass>;
+  }
+}
+
+multiclass VPatConversionVI_WF_RM_BF16<string intrinsic, string instruction> {
+  foreach vtiToWti = AllWidenableIntToBF16Vectors in {
+    defvar vti = vtiToWti.Vti;
+    defvar fwti = vtiToWti.Wti;
+    let Predicates = !listconcat(GetVTypePredicates<vti>.Predicates,
+                                 GetVTypePredicates<fwti>.Predicates) in
+    defm : VPatConversionRoundingMode<intrinsic, instruction, "W",
+                                      vti.Vector, fwti.Vector, vti.Mask, vti.Log2SEW,
+                                      vti.LMul, vti.RegClass, fwti.RegClass>;
+  }
+}
+
+multiclass VPatConversionVF_WF_BF16<string intrinsic, string instruction,
+                                    bit isSEWAware = 0> {
+  foreach fvtiToFWti = AllWidenableBF16ToFloatVectors in {
+    defvar fvti = fvtiToFWti.Vti;
+    defvar fwti = fvtiToFWti.Wti;
+    let Predicates = !listconcat(GetVTypePredicates<fvti>.Predicates,
+                                 GetVTypePredicates<fwti>.Predicates) in
+    defm : VPatConversion<intrinsic, instruction, "W",
+                          fvti.Vector, fwti.Vector, fvti.Mask, fvti.Log2SEW,
+                          fvti.LMul, fvti.RegClass, fwti.RegClass, isSEWAware>;
+  }
+}
+
+let Predicates = [HasStdExtZvfbfa] in {
+defm : VPatBinaryV_VV_VX_RM<"int_riscv_vfadd", "PseudoVFADD_ALT",
+                            AllBF16Vectors, isSEWAware = 1>;
+defm : VPatBinaryV_VV_VX_RM<"int_riscv_vfsub", "PseudoVFSUB_ALT",
+                            AllBF16Vectors, isSEWAware = 1>;
+defm : VPatBinaryV_VX_RM<"int_riscv_vfrsub", "PseudoVFRSUB_ALT",
+                         AllBF16Vectors, isSEWAware = 1>;
+defm : VPatBinaryW_VV_VX_RM<"int_riscv_vfwadd", "PseudoVFWADD_ALT",
+                            AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatBinaryW_VV_VX_RM<"int_riscv_vfwsub", "PseudoVFWSUB_ALT",
+                            AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatBinaryW_WV_WX_RM<"int_riscv_vfwadd_w", "PseudoVFWADD_ALT",
+                            AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatBinaryW_WV_WX_RM<"int_riscv_vfwsub_w", "PseudoVFWSUB_ALT",
+                            AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX_RM<"int_riscv_vfmul", "PseudoVFMUL_ALT",
+                            AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryW_VV_VX_RM<"int_riscv_vfwmul", "PseudoVFWMUL_ALT",
+                            AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfmacc", "PseudoVFMACC_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfnmacc", "PseudoVFNMACC_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfmsac", "PseudoVFMSAC_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfnmsac", "PseudoVFNMSAC_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfmadd", "PseudoVFMADD_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfnmadd", "PseudoVFNMADD_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfmsub", "PseudoVFMSUB_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryV_VV_VX_AAXA_RM<"int_riscv_vfnmsub", "PseudoVFNMSUB_ALT",
+                                  AllBF16Vectors, isSEWAware=1>;
+defm : VPatTernaryW_VV_VX_RM<"int_riscv_vfwmacc", "PseudoVFWMACC_ALT",
+                             AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatTernaryW_VV_VX_RM<"int_riscv_vfwnmacc", "PseudoVFWNMACC_ALT",
+                             AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatTernaryW_VV_VX_RM<"int_riscv_vfwmsac", "PseudoVFWMSAC_ALT",
+                             AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatTernaryW_VV_VX_RM<"int_riscv_vfwnmsac", "PseudoVFWNMSAC_ALT",
+                             AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatUnaryV_V<"int_riscv_vfrsqrt7", "PseudoVFRSQRT7_ALT",
+                    AllBF16Vectors, isSEWAware=1>;
+defm : VPatUnaryV_V_RM<"int_riscv_vfrec7", "PseudoVFREC7_ALT",
+                       AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX<"int_riscv_vfmin", "PseudoVFMIN_ALT",
+                         AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX<"int_riscv_vfmax", "PseudoVFMAX_ALT",
+                         AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX<"int_riscv_vfsgnj", "PseudoVFSGNJ_ALT",
+                         AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX<"int_riscv_vfsgnjn", "PseudoVFSGNJN_ALT",
+                         AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryV_VV_VX<"int_riscv_vfsgnjx", "PseudoVFSGNJX_ALT",
+                         AllBF16Vectors, isSEWAware=1>;
+defm : VPatBinaryM_VV_VX<"int_riscv_vmfeq", "PseudoVMFEQ_ALT", AllBF16Vectors>;
+defm : VPatBinaryM_VV_VX<"int_riscv_vmfle", "PseudoVMFLE_ALT", AllBF16Vectors>;
+defm : VPatBinaryM_VV_VX<"int_riscv_vmflt", "PseudoVMFLT_ALT", AllBF16Vectors>;
+defm : VPatBinaryM_VV_VX<"int_riscv_vmfne", "PseudoVMFNE_ALT", AllBF16Vectors>;
+defm : VPatBinaryM_VX<"int_riscv_vmfgt", "PseudoVMFGT_ALT", AllBF16Vectors>;
+defm : VPatBinaryM_VX<"int_riscv_vmfge", "PseudoVMFGE_ALT", AllBF16Vectors>;
+defm : VPatBinarySwappedM_VV<"int_riscv_vmfgt", "PseudoVMFLT_ALT", AllBF16Vectors>;
+defm : VPatBinarySwappedM_VV<"int_riscv_vmfge", "PseudoVMFLE_ALT", AllBF16Vectors>;
+defm : VPatConversionVI_VF_BF16<"int_riscv_vfclass", "PseudoVFCLASS_ALT">;
+foreach vti = AllBF16Vectors in {
+  let Predicates = GetVTypePredicates<vti>.Predicates in
+    defm : VPatBinaryCarryInTAIL<"int_riscv_vfmerge", "PseudoVFMERGE_ALT",
+                                 "V"#vti.ScalarSuffix#"M",
+                                 vti.Vector,
+                                 vti.Vector, vti.Scalar, vti.Mask,
+                                 vti.Log2SEW, vti.LMul, vti.RegClass,
+                                 vti.RegClass, vti.ScalarRegClass>;
+}
+defm : VPatConversionWF_VI_BF16<"int_riscv_vfwcvt_f_xu_v", "PseudoVFWCVT_F_XU_ALT",
+                                isSEWAware=1>;
+defm : VPatConversionWF_VI_BF16<"int_riscv_vfwcvt_f_x_v", "PseudoVFWCVT_F_X_ALT",
+                                isSEWAware=1>;
+defm : VPatConversionWF_VF_BF16<"int_riscv_vfwcvt_f_f_v", "PseudoVFWCVT_F_F_ALT",
+                                isSEWAware=1>;
+defm : VPatConversionVI_WF_RM_BF16<"int_riscv_vfncvt_xu_f_w", "PseudoVFNCVT_XU_F_ALT">;
+defm : VPatConversionVI_WF_RM_BF16<"int_riscv_vfncvt_x_f_w", "PseudoVFNCVT_X_F_ALT">;
+defm : VPatConversionVI_WF_BF16<"int_riscv_vfncvt_rtz_xu_f_w", "PseudoVFNCVT_RTZ_XU_F_ALT">;
+defm : VPatConversionVI_WF_BF16<"int_riscv_vfncvt_rtz_x_f_w", "PseudoVFNCVT_RTZ_X_F_ALT">;
+defm : VPatConversionVF_WF_RM<"int_riscv_vfncvt_f_f_w", "PseudoVFNCVT_F_F_ALT",
+                              AllWidenableBF16ToFloatVectors, isSEWAware=1>;
+defm : VPatConversionVF_WF_BF16<"int_riscv_vfncvt_rod_f_f_w", "PseudoVFNCVT_ROD_F_F_ALT",
+                                isSEWAware=1>;
+defm : VPatBinaryV_VX<"int_riscv_vfslide1up", "PseudoVFSLIDE1UP_ALT", AllBF16Vectors>;
+defm : VPatBinaryV_VX<"int_riscv_vfslide1down", "PseudoVFSLIDE1DOWN_ALT", AllBF16Vectors>;
+
+foreach fvti = AllBF16Vectors in {
+  defvar ivti = GetIntVTypeInfo<fvti>.Vti;
+  let Predicates = GetVTypePredicates<ivti>.Predicates in {
+    // 13.16. Vector Floating-Point Move Instruction
+    // If we're splatting fpimm0, use vmv.v.x vd, x0.
+    def : Pat<(fvti.Vector (riscv_vfmv_v_f_vl
+                           fvti.Vector:$passthru, (fvti.Scalar (fpimm0)), VLOpFrag)),
+              (!cast<Instruction>("PseudoVMV_V_I_"#fvti.LMul.MX)
+               $passthru, 0, GPR:$vl, fvti.Log2SEW, TU_MU)>;
+    def : Pat<(fvti.Vector (riscv_vfmv_v_f_vl
+                           fvti.Vector:$passthru, (fvti.Scalar (SelectScalarFPAsInt (XLenVT GPR:$imm))), VLOpFrag)),
+              (!cast<Instruction>("PseudoVMV_V_X_"#fvti.LMul.MX)
+               $passthru, GPR:$imm, GPR:$vl, fvti.Log2SEW, TU_MU)>;
+  }
+
+  let Predicates = GetVTypePredicates<fvti>.Predicates in {
+    def : Pat<(fvti.Vector (riscv_vfmv_v_f_vl
+                           fvti.Vector:$passthru, (fvti.Scalar fvti.ScalarRegClass:$rs2), VLOpFrag)),
+              (!cast<Instruction>("PseudoVFMV_V_ALT_" # fvti.ScalarSuffix # "_" #
+                                  fvti.LMul.MX)
+               $passthru, (fvti.Scalar fvti.ScalarRegClass:$rs2),
+               GPR:$vl, fvti.Log2SEW, TU_MU)>;
+  }
+}
+
+foreach vti = NoGroupBF16Vectors in {
+  let Predicates = GetVTypePredicates<vti>.Predicates in {
+    def : Pat<(vti.Vector (riscv_vfmv_s_f_vl (vti.Vector vti.RegClass:$passthru),
+                                             (vti.Scalar (fpimm0)),
+                                             VLOpFrag)),
+              (PseudoVMV_S_X $passthru, (XLenVT X0), GPR:$vl, vti.Log2SEW)>;
+    def : Pat<(vti.Vector (riscv_vfmv_s_f_vl (vti.Vector vti.RegClass:$passthru),
+                                             (vti.Scalar (SelectScalarFPAsInt (XLenVT GPR:$imm))),
+                                             VLOpFrag)),
+              (PseudoVMV_S_X $passthru, GPR:$imm, GPR:$vl, vti.Log2SEW)>;
+    def : Pat<(vti.Vector (riscv_vfmv_s_f_vl (vti.Vector vti.RegClass:$passthru),
+                                             vti.ScalarRegClass:$rs1,
+                                             VLOpFrag)),
+              (!cast<Instruction>("PseudoVFMV_S_"#vti.ScalarSuffix#"_ALT")
+                  vti.RegClass:$passthru,
+                  (vti.Scalar vti.ScalarRegClass:$rs1), GPR:$vl, vti.Log2SEW)>;
+  }
+
+  defvar vfmv_f_s_inst = !cast<Instruction>(!strconcat("PseudoVFMV_",
+                                                       vti.ScalarSuffix,
+                                                       "_S_ALT"));
+  // Only pattern-match extract-element operations where the index is 0. Any
+  // other index will have been custom-lowered to slide the vector correctly
+  // into place.
+  let Predicates = GetVTypePredicates<vti>.Predicates in
+  def : Pat<(vti.Scalar (extractelt (vti.Vector vti.RegClass:$rs2), 0)),
+            (vfmv_f_s_inst vti.RegClass:$rs2, vti.Log2SEW)>;
+}
+} // Predicates = [HasStdExtZvfbfa]
diff --git a/llvm/lib/Target/RISCV/RISCVSubtarget.h b/llvm/lib/Target/RISCV/RISCVSubtarget.h
index 7dffa63d85505..a58d138f3e230 100644
--- a/llvm/lib/Target/RISCV/RISCVSubtarget.h
+++ b/llvm/lib/Target/RISCV/RISCVSubtarget.h
@@ -288,9 +288,12 @@ class RISCVSubtarget : public RISCVGenSubtargetInfo {
   bool hasVInstructionsI64() const { return HasStdExtZve64x; }
   bool hasVInstructionsF16Minimal() const { return HasStdExtZvfhmin; }
   bool hasVInstructionsF16() const { return HasStdExtZvfh; }
-  bool hasVInstructionsBF16Minimal() const { return HasStdExtZvfbfmin; }
+  bool hasVInstructionsBF16Minimal() const {
+    return HasStdExtZvfbfmin || HasStdExtZvfbfa;
+  }
   bool hasVInstructionsF32() const { return HasStdExtZve32f; }
   bool hasVInstructionsF64() const { return HasStdExtZve64d; }
+  bool hasVInstructionsBF16() const { return HasStdExtZvfbfa; }
   // F16 and F64 both require F32.
   bool hasVInstructionsAnyF() const { return hasVInstructionsF32(); }
   bool hasVInstructionsFullMultiply() const { return HasStdExtV; }
diff --git a/llvm/test/CodeGen/RISCV/rvv/mixed-float-bf16-arith.ll b/llvm/test/CodeGen/RISCV/rvv/mixed-float-bf16-arith.ll
new file mode 100644
index 0000000000000..145fc5dee4b4c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/mixed-float-bf16-arith.ll
@@ -0,0 +1,186 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+zvfh,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+zvfh,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x half> @llvm.riscv.vfadd.nxv1f16.nxv1f16(
+  <vscale x 1 x half>,
+  <vscale x 1 x half>,
+  <vscale x 1 x half>,
+  iXLen, iXLen);
+
+declare <vscale x 1 x i32> @llvm.riscv.vadd.nxv1i32.nxv1i32(
+  <vscale x 1 x i32>,
+  <vscale x 1 x i32>,
+  <vscale x 1 x i32>,
+  iXLen);
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @test_half_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2, <vscale x 1 x half> %3, <vscale x 1 x half> %4, ptr %ptr) nounwind {
+; CHECK-LABEL: test_half_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a2, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v10, v10, v11
+; CHECK-NEXT:    vsetvli zero, zero, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a2
+; CHECK-NEXT:    vse16.v v10, (a1)
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x half> @llvm.riscv.vfadd.nxv1f16.nxv1f16(
+    <vscale x 1 x half> poison,
+    <vscale x 1 x half> %3,
+    <vscale x 1 x half> %4,
+    iXLen 0, iXLen %2)
+
+  %b = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  call void @llvm.riscv.vse(<vscale x 1 x half> %a, ptr %ptr, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %b
+}
+
+define <vscale x 1 x bfloat> @test_i32_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2, <vscale x 1 x i32> %3, <vscale x 1 x i32> %4, ptr %ptr) nounwind {
+; CHECK-LABEL: test_i32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e32, mf2, ta, ma
+; CHECK-NEXT:    vadd.vv v10, v10, v11
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vsetvli zero, zero, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    vse32.v v10, (a1)
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i32> @llvm.riscv.vadd.nxv1i32.nxv1i32(
+    <vscale x 1 x i32> poison,
+    <vscale x 1 x i32> %3,
+    <vscale x 1 x i32> %4,
+    iXLen %2)
+
+  %b = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  call void @llvm.riscv.vse(<vscale x 1 x i32> %a, ptr %ptr, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %b
+}
+
+define <vscale x 1 x bfloat> @test_half_bf16_half(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2, <vscale x 1 x half> %3, <vscale x 1 x half> %4, ptr %ptr) nounwind {
+; CHECK-LABEL: test_half_bf16_half:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a2, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v10, v10, v11
+; CHECK-NEXT:    vsetvli zero, zero, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v9, v10, v11
+; CHECK-NEXT:    fsrm a2
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vse16.v v9, (a1)
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x half> @llvm.riscv.vfadd.nxv1f16.nxv1f16(
+    <vscale x 1 x half> poison,
+    <vscale x 1 x half> %3,
+    <vscale x 1 x half> %4,
+    iXLen 0, iXLen %2)
+
+  %b = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  %c = call <vscale x 1 x half> @llvm.riscv.vfadd.nxv1f16.nxv1f16(
+    <vscale x 1 x half> poison,
+    <vscale x 1 x half> %a,
+    <vscale x 1 x half> %4,
+    iXLen 0, iXLen %2)
+
+  store <vscale x 1 x half> %c, ptr %ptr
+
+  ret <vscale x 1 x bfloat> %b
+}
+
+define <vscale x 1 x bfloat> @test_bf16_half_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2, <vscale x 1 x half> %3, <vscale x 1 x half> %4, ptr %ptr) nounwind {
+; CHECK-LABEL: test_bf16_half_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a2, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    vsetvli zero, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v10, v10, v11
+; CHECK-NEXT:    vsetvli zero, zero, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a2
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vse16.v v10, (a1)
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  %b = call <vscale x 1 x half> @llvm.riscv.vfadd.nxv1f16.nxv1f16(
+    <vscale x 1 x half> poison,
+    <vscale x 1 x half> %3,
+    <vscale x 1 x half> %4,
+    iXLen 0, iXLen %2)
+
+  %c = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %a,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  store <vscale x 1 x half> %b, ptr %ptr
+
+  ret <vscale x 1 x bfloat> %c
+}
+
+define <vscale x 1 x bfloat> @test_bf16_i16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2, <vscale x 1 x i16> %3, <vscale x 1 x i16> %4, ptr %ptr) nounwind {
+; CHECK-LABEL: test_bf16_i16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a2, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    vadd.vv v9, v10, v11
+; CHECK-NEXT:    fsrm a2
+; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
+; CHECK-NEXT:    vse16.v v9, (a1)
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  %b = call <vscale x 1 x i16> @llvm.riscv.vadd.nxv1i16.nxv1i16(
+    <vscale x 1 x i16> poison,
+    <vscale x 1 x i16> %3,
+    <vscale x 1 x i16> %4,
+    iXLen %2)
+
+  store <vscale x 1 x i16> %b, ptr %ptr
+
+  ret <vscale x 1 x bfloat> %a
+}
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfadd-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfadd-bf.ll
new file mode 100644
index 0000000000000..db1b081258d5f
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfadd-bf.ll
@@ -0,0 +1,607 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfadd_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfadd_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfadd_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfadd_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfadd_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfadd_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfadd.vv v8, v8, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfadd_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfadd.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfadd_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfadd_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfadd.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfadd_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfadd_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfadd.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfadd_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfadd_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfadd.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfadd_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfadd_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfadd.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfadd_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfadd_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfadd.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfadd_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfadd.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfadd_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfadd_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfadd.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfadd.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfclass-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfclass-bf.ll
new file mode 100644
index 0000000000000..d7d49b379b5a4
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfclass-bf.ll
@@ -0,0 +1,294 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1i16.nxv1bf16(
+  <vscale x 1 x i16>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i16> @intrinsic_vfclass_v_nxv1i16_nxv1bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv1i16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 1 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 1 x i16> @llvm.riscv.vfclass.nxv1i16.nxv1bf16(
+    <vscale x 1 x i16> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x i16> %a
+}
+
+declare <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1i16.nxv1bf16(
+  <vscale x 1 x i16>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 1 x i16> @intrinsic_vfclass_mask_v_nxv1i16_nxv1bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv1i16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 1 x i16> %0,
+  <vscale x 1 x bfloat> %1,
+  <vscale x 1 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 1 x i16> @llvm.riscv.vfclass.mask.nxv1i16.nxv1bf16(
+    <vscale x 1 x i16> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x i16> %a
+}
+
+declare <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2i16.nxv2bf16(
+  <vscale x 2 x i16>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i16> @intrinsic_vfclass_v_nxv2i16_nxv2bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv2i16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 2 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 2 x i16> @llvm.riscv.vfclass.nxv2i16.nxv2bf16(
+    <vscale x 2 x i16> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x i16> %a
+}
+
+declare <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2i16.nxv2bf16(
+  <vscale x 2 x i16>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 2 x i16> @intrinsic_vfclass_mask_v_nxv2i16_nxv2bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv2i16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 2 x i16> %0,
+  <vscale x 2 x bfloat> %1,
+  <vscale x 2 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 2 x i16> @llvm.riscv.vfclass.mask.nxv2i16.nxv2bf16(
+    <vscale x 2 x i16> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4i16.nxv4bf16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vfclass_v_nxv4i16_nxv4bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv4i16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 4 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.vfclass.nxv4i16.nxv4bf16(
+    <vscale x 4 x i16> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4i16.nxv4bf16(
+  <vscale x 4 x i16>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 4 x i16> @intrinsic_vfclass_mask_v_nxv4i16_nxv4bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv4i16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 4 x i16> %0,
+  <vscale x 4 x bfloat> %1,
+  <vscale x 4 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 4 x i16> @llvm.riscv.vfclass.mask.nxv4i16.nxv4bf16(
+    <vscale x 4 x i16> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8i16.nxv8bf16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vfclass_v_nxv8i16_nxv8bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv8i16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 8 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.vfclass.nxv8i16.nxv8bf16(
+    <vscale x 8 x i16> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8i16.nxv8bf16(
+  <vscale x 8 x i16>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 8 x i16> @intrinsic_vfclass_mask_v_nxv8i16_nxv8bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv8i16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v10, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 8 x i16> %0,
+  <vscale x 8 x bfloat> %1,
+  <vscale x 8 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 8 x i16> @llvm.riscv.vfclass.mask.nxv8i16.nxv8bf16(
+    <vscale x 8 x i16> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16i16.nxv16bf16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vfclass_v_nxv16i16_nxv16bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv16i16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 16 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.vfclass.nxv16i16.nxv16bf16(
+    <vscale x 16 x i16> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16i16.nxv16bf16(
+  <vscale x 16 x i16>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 16 x i16> @intrinsic_vfclass_mask_v_nxv16i16_nxv16bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv16i16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v12, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 16 x i16> %0,
+  <vscale x 16 x bfloat> %1,
+  <vscale x 16 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 16 x i16> @llvm.riscv.vfclass.mask.nxv16i16.nxv16bf16(
+    <vscale x 16 x i16> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32i16.nxv32bf16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vfclass_v_nxv32i16_nxv32bf16(
+; CHECK-LABEL: intrinsic_vfclass_v_nxv32i16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfclass.v v8, v8
+; CHECK-NEXT:    ret
+  <vscale x 32 x bfloat> %0,
+  iXLen %1) nounwind {
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.vfclass.nxv32i16.nxv32bf16(
+    <vscale x 32 x i16> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x i16> %a
+}
+
+declare <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32i16.nxv32bf16(
+  <vscale x 32 x i16>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 32 x i16> @intrinsic_vfclass_mask_v_nxv32i16_nxv32bf16(
+; CHECK-LABEL: intrinsic_vfclass_mask_v_nxv32i16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, tu, mu
+; CHECK-NEXT:    vfclass.v v8, v16, v0.t
+; CHECK-NEXT:    ret
+  <vscale x 32 x i16> %0,
+  <vscale x 32 x bfloat> %1,
+  <vscale x 32 x i1> %2,
+  iXLen %3) nounwind {
+entry:
+  %a = call <vscale x 32 x i16> @llvm.riscv.vfclass.mask.nxv32i16.nxv32bf16(
+    <vscale x 32 x i16> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3, iXLen 0)
+
+  ret <vscale x 32 x i16> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmacc-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmacc-bf.ll
new file mode 100644
index 0000000000000..0937a82b48580
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmacc-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmacc_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmacc_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmacc_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmacc_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmacc_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmacc_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmacc_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmacc.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmacc_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmacc.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmacc_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmacc.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmacc_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmacc.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmacc_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmacc_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmacc.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmacc_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmacc_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmacc.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmacc_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmacc_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmacc.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmacc_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmacc_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmacc.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmacc_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmacc_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmacc_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmacc.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmacc.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmadd-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmadd-bf.ll
new file mode 100644
index 0000000000000..795fd4f56ec2c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmadd-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmadd_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmadd_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmadd_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmadd_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmadd.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmadd.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmadd_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmadd.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmadd.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmadd_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmadd_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmadd.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmadd_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmadd_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmadd.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmadd_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmadd_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmadd.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmadd_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmadd_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmadd.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmadd_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmadd_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmadd_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmadd.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmadd.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmax-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmax-bf.ll
new file mode 100644
index 0000000000000..a337d3061ce78
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmax-bf.ll
@@ -0,0 +1,571 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmax_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmax_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmax_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmax_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmax_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmax_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmax_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmax_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmax_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmax_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmax_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmax.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmax_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmax.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmax_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmax_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmax.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmax_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmax_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmax.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmax_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmax_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmax.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmax_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmax_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmax.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmax_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmax_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmax.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmax_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmax.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmax_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmax_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmax.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmax.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmerge-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmerge-bf.ll
new file mode 100644
index 0000000000000..86ba7c7fb7fe6
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmerge-bf.ll
@@ -0,0 +1,258 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmerge_vfm_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmerge_vfm_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmerge_vfm_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmerge_vfm_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmerge_vfm_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmerge_vfm_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vfm_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa0, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+define <vscale x 1 x bfloat> @intrinsic_vfmerge_vzm_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmerge.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 1 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+define <vscale x 2 x bfloat> @intrinsic_vfmerge_vzm_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmerge.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 2 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+define <vscale x 4 x bfloat> @intrinsic_vfmerge_vzm_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmerge.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 4 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+define <vscale x 8 x bfloat> @intrinsic_vfmerge_vzm_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmerge.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 8 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+define <vscale x 16 x bfloat> @intrinsic_vfmerge_vzm_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmerge.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 16 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+define <vscale x 32 x bfloat> @intrinsic_vfmerge_vzm_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x i1> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmerge_vzm_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fmv.h.x fa5, zero
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmerge.vfm v8, v8, fa5, v0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmerge.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat zeroinitializer,
+    <vscale x 32 x i1> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmin-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmin-bf.ll
new file mode 100644
index 0000000000000..37c0cf506a6fa
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmin-bf.ll
@@ -0,0 +1,571 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmin_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmin_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmin_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmin_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmin_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmin_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmin_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmin_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmin_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmin_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmin_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmin.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmin_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmin.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmin_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmin_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmin.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmin_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmin_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmin.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmin_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmin_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmin.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmin_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmin_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmin.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmin_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmin_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmin.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmin_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmin.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmin_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmin_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmin.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmin.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmsac-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmsac-bf.ll
new file mode 100644
index 0000000000000..ebeda9eaf42c2
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmsac-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsac_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsac_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsac_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsac_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsac_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsac_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsac_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmsac.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsac_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmsac.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsac_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmsac.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsac_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmsac.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsac_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmsac_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsac.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsac_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmsac_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsac.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsac_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmsac_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsac.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsac_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmsac_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsac.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsac_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmsac_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsac_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmsac.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsac.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmsub-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmsub-bf.ll
new file mode 100644
index 0000000000000..b032c1d66f3b9
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmsub-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsub_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsub_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsub_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsub_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmsub.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmsub.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsub_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmsub.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmsub.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfmsub_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmsub_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmsub.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfmsub_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmsub_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmsub.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfmsub_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmsub_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmsub.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfmsub_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmsub_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmsub.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfmsub_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmsub_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmsub_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfmsub.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmsub.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmul-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmul-bf.ll
new file mode 100644
index 0000000000000..44bce723c39d4
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmul-bf.ll
@@ -0,0 +1,607 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmul_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmul_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmul_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmul_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmul_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmul_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmul_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmul_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmul_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmul_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmul_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmul.vv v8, v8, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmul_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmul.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmul_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmul_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmul.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmul_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmul_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmul.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmul_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmul_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmul.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmul_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmul_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmul.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmul_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmul_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmul.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmul_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmul.vf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmul_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfmul_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfmul.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmul.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmv-bf-s.ll b/llvm/test/CodeGen/RISCV/rvv/vfmv-bf-s.ll
new file mode 100644
index 0000000000000..fbc7311945c8b
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmv-bf-s.ll
@@ -0,0 +1,88 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+d,+v,+experimental-zvfbfa -target-abi lp64d -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv32 -mattr=+d,+v,+experimental-zvfbfa -target-abi ilp32d -verify-machineinstrs < %s | FileCheck %s
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv1bf16(<vscale x 1 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv1bf16(<vscale x 1 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv1bf16(<vscale x 1 x bfloat> %0)
+  ret bfloat %a
+}
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv2bf16(<vscale x 2 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv2bf16(<vscale x 2 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv2bf16(<vscale x 2 x bfloat> %0)
+  ret bfloat %a
+}
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv4bf16(<vscale x 4 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv4bf16(<vscale x 4 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv4bf16(<vscale x 4 x bfloat> %0)
+  ret bfloat %a
+}
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv8bf16(<vscale x 8 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv8bf16(<vscale x 8 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv8bf16(<vscale x 8 x bfloat> %0)
+  ret bfloat %a
+}
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv16bf16(<vscale x 16 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv16bf16(<vscale x 16 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv16bf16(<vscale x 16 x bfloat> %0)
+  ret bfloat %a
+}
+
+declare bfloat @llvm.riscv.vfmv.f.s.nxv32bf16(<vscale x 32 x bfloat>)
+
+define bfloat @intrinsic_vfmv.f.s_s_nxv32bf16(<vscale x 32 x bfloat> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetivli zero, 1, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.x.s a0, v8
+; CHECK-NEXT:    fmv.h.x fa0, a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call bfloat @llvm.riscv.vfmv.f.s.nxv32bf16(<vscale x 32 x bfloat> %0)
+  ret bfloat %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmv-s-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmv-s-bf.ll
new file mode 100644
index 0000000000000..a810809fca515
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmv-s-bf.ll
@@ -0,0 +1,161 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s --check-prefixes=CHECK
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s --check-prefixes=CHECK
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16(<vscale x 1 x bfloat>, bfloat, iXLen)
+
+define <vscale x 1 x bfloat> @intrinsic_vfmv.s.f_f_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16(<vscale x 2 x bfloat>, bfloat, iXLen)
+
+define <vscale x 2 x bfloat> @intrinsic_vfmv.s.f_f_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16(<vscale x 4 x bfloat>, bfloat, iXLen)
+
+define <vscale x 4 x bfloat> @intrinsic_vfmv.s.f_f_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16(<vscale x 8 x bfloat>, bfloat, iXLen)
+
+define <vscale x 8 x bfloat> @intrinsic_vfmv.s.f_f_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16(<vscale x 16 x bfloat>, bfloat, iXLen)
+
+define <vscale x 16 x bfloat> @intrinsic_vfmv.s.f_f_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16(<vscale x 32 x bfloat>, bfloat, iXLen)
+
+define <vscale x 32 x bfloat> @intrinsic_vfmv.s.f_f_nxv32bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfmv.s.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2)
+  ret <vscale x 32 x bfloat> %a
+}
+
+define <vscale x 1 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16(<vscale x 1 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 1 x bfloat> %a
+}
+
+define <vscale x 2 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.s.f.nxv2bf16(<vscale x 2 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 2 x bfloat> %a
+}
+
+define <vscale x 4 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.s.f.nxv4bf16(<vscale x 4 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 4 x bfloat> %a
+}
+
+define <vscale x 8 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.s.f.nxv8bf16(<vscale x 8 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 8 x bfloat> %a
+}
+
+define <vscale x 16 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.s.f.nxv16bf16(<vscale x 16 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 16 x bfloat> %a
+}
+
+define <vscale x 32 x bfloat> @intrinsic_vfmv.s.f_f_zero_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_zero_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, zero
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.s.f.nxv32bf16(<vscale x 32 x bfloat> %0, bfloat 0.0, iXLen %1)
+  ret <vscale x 32 x bfloat> %a
+}
+
+define <vscale x 1 x bfloat> @intrinsic_vfmv.s.f_f_nxv1bf16_negzero(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1bf16_negzero:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    lui a1, 1048568
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, tu, ma
+; CHECK-NEXT:    vmv.s.x v8, a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.s.f.nxv1bf16(<vscale x 1 x bfloat> %0, bfloat -0.0, iXLen %1)
+  ret <vscale x 1 x bfloat> %a
+}
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfmv-v-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfmv-v-bf.ll
new file mode 100644
index 0000000000000..f3293ddc83ef9
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfmv-v-bf.ll
@@ -0,0 +1,216 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfmv.v.f_f_nxv1bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfmv.v.f_f_nxv2bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfmv.v.f_f_nxv4bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfmv.v.f_f_nxv8bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfmv.v.f_f_nxv16bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfmv.v.f_f_nxv32bf16(bfloat %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_f_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfmv.v.f v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    bfloat %0,
+    iXLen %1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+define <vscale x 1 x bfloat> @intrinsic_vfmv.v.f_zero_nxv1bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.v.f_zero_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfmv.v.f.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+define <vscale x 2 x bfloat> @intrinsic_vmv.v.i_zero_nxv2bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vmv.v.i_zero_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, mf2, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfmv.v.f.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+define <vscale x 4 x bfloat> @intrinsic_vmv.v.i_zero_nxv4bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vmv.v.i_zero_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m1, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfmv.v.f.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+define <vscale x 8 x bfloat> @intrinsic_vmv.v.i_zero_nxv8bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vmv.v.i_zero_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m2, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfmv.v.f.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+define <vscale x 16 x bfloat> @intrinsic_vmv.v.i_zero_nxv16bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vmv.v.i_zero_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m4, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfmv.v.f.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+define <vscale x 32 x bfloat> @intrinsic_vmv.v.i_zero_nxv32bf16(iXLen %0) nounwind {
+; CHECK-LABEL: intrinsic_vmv.v.i_zero_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16, m8, ta, ma
+; CHECK-NEXT:    vmv.v.i v8, 0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfmv.v.f.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    bfloat 0.0,
+    iXLen %0)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfncvt-rod-bf-f.ll b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rod-bf-f.ll
new file mode 100644
index 0000000000000..7d587fd55cd83
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rod-bf-f.ll
@@ -0,0 +1,226 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x float>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfncvt_rod.f.f.w_nxv1bf16_nxv1f32(<vscale x 1 x float> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rod.f.f.w_nxv1bf16_nxv1f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfncvt.rod.f.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv1bf16.nxv1f32(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x float> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x float>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfncvt_mask_rod.f.f.w_nxv1bf16_nxv1f32(<vscale x 1 x bfloat> %0, <vscale x 1 x float> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rod.f.f.w_nxv1bf16_nxv1f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfncvt.rod.f.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv1bf16.nxv1f32(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x float> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x float>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfncvt_rod.f.f.w_nxv2bf16_nxv2f32(<vscale x 2 x float> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rod.f.f.w_nxv2bf16_nxv2f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfncvt.rod.f.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv2bf16.nxv2f32(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x float> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x float>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfncvt_mask_rod.f.f.w_nxv2bf16_nxv2f32(<vscale x 2 x bfloat> %0, <vscale x 2 x float> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rod.f.f.w_nxv2bf16_nxv2f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfncvt.rod.f.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv2bf16.nxv2f32(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x float> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x float>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfncvt_rod.f.f.w_nxv4bf16_nxv4f32(<vscale x 4 x float> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rod.f.f.w_nxv4bf16_nxv4f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfncvt.rod.f.f.w v10, v8
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv4bf16.nxv4f32(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x float> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x float>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfncvt_mask_rod.f.f.w_nxv4bf16_nxv4f32(<vscale x 4 x bfloat> %0, <vscale x 4 x float> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rod.f.f.w_nxv4bf16_nxv4f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfncvt.rod.f.f.w v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv4bf16.nxv4f32(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x float> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x float>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfncvt_rod.f.f.w_nxv8bf16_nxv8f32(<vscale x 8 x float> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rod.f.f.w_nxv8bf16_nxv8f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfncvt.rod.f.f.w v12, v8
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv8bf16.nxv8f32(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x float> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x float>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfncvt_mask_rod.f.f.w_nxv8bf16_nxv8f32(<vscale x 8 x bfloat> %0, <vscale x 8 x float> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rod.f.f.w_nxv8bf16_nxv8f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfncvt.rod.f.f.w v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv8bf16.nxv8f32(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x float> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x float>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfncvt_rod.f.f.w_nxv16bf16_nxv16f32(<vscale x 16 x float> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rod.f.f.w_nxv16bf16_nxv16f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfncvt.rod.f.f.w v16, v8
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.nxv16bf16.nxv16f32(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x float> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x float>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfncvt_mask_rod.f.f.w_nxv16bf16_nxv16f32(<vscale x 16 x bfloat> %0, <vscale x 16 x float> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rod.f.f.w_nxv16bf16_nxv16f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfncvt.rod.f.f.w v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfncvt.rod.f.f.w.mask.nxv16bf16.nxv16f32(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x float> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-x-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-x-bf.ll
new file mode 100644
index 0000000000000..ee9e3d1b9f630
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-x-bf.ll
@@ -0,0 +1,270 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv1i8_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv1i8_nxv1bf16(<vscale x 1 x i8> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv2i8_nxv2bf16(<vscale x 2 x i8> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv4i8_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv4i8_nxv4bf16(<vscale x 4 x i8> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv8i8_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v10, v8
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv8i8_nxv8bf16(<vscale x 8 x i8> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv16i8_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v12, v8
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv16i8_nxv16bf16(<vscale x 16 x i8> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_rtz.x.f.w_nxv32i8_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.x.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v16, v8
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_mask_rtz.x.f.w_nxv32i8_nxv32bf16(<vscale x 32 x i8> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.x.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.x.f.w v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.x.f.w.mask.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x i8> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-xu-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-xu-bf.ll
new file mode 100644
index 0000000000000..521f7274dc5c9
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfncvt-rtz-xu-bf.ll
@@ -0,0 +1,270 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv1i8_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv1i8_nxv1bf16(<vscale x 1 x i8> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv2i8_nxv2bf16(<vscale x 2 x i8> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv4i8_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv4i8_nxv4bf16(<vscale x 4 x i8> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv8i8_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v10, v8
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv8i8_nxv8bf16(<vscale x 8 x i8> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv16i8_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v12, v8
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv16i8_nxv16bf16(<vscale x 16 x i8> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_rtz.xu.f.w_nxv32i8_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_rtz.xu.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v16, v8
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_mask_rtz.xu.f.w_nxv32i8_nxv32bf16(<vscale x 32 x i8> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_rtz.xu.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfncvt.rtz.xu.f.w v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.rtz.xu.f.w.mask.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x i8> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfncvt-x-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfncvt-x-bf.ll
new file mode 100644
index 0000000000000..ab9ebade287e6
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfncvt-x-bf.ll
@@ -0,0 +1,288 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_x.f.w_nxv1i8_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv1i8_nxv1bf16(<vscale x 1 x i8> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_x.f.w_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv2i8_nxv2bf16(<vscale x 2 x i8> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_x.f.w_nxv4i8_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv4i8_nxv4bf16(<vscale x 4 x i8> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_x.f.w_nxv8i8_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v10, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv8i8_nxv8bf16(<vscale x 8 x i8> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_x.f.w_nxv16i8_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v12, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv16i8_nxv16bf16(<vscale x 16 x i8> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_x.f.w_nxv32i8_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_x.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vfncvt.x.f.w v16, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_mask_x.f.w_nxv32i8_nxv32bf16(<vscale x 32 x i8> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_x.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfncvt.x.f.w v8, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.x.f.w.mask.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x i8> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfncvt-xu-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfncvt-xu-bf.ll
new file mode 100644
index 0000000000000..61c6803ce12bd
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfncvt-xu-bf.ll
@@ -0,0 +1,288 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_xu.f.w_nxv1i8_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16(
+  <vscale x 1 x i8>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv1i8_nxv1bf16(<vscale x 1 x i8> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv1i8_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv1i8.nxv1bf16(
+    <vscale x 1 x i8> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_xu.f.w_nxv2i8_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16(
+  <vscale x 2 x i8>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv2i8_nxv2bf16(<vscale x 2 x i8> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv2i8_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv2i8.nxv2bf16(
+    <vscale x 2 x i8> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_xu.f.w_nxv4i8_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16(
+  <vscale x 4 x i8>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv4i8_nxv4bf16(<vscale x 4 x i8> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv4i8_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv4i8.nxv4bf16(
+    <vscale x 4 x i8> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_xu.f.w_nxv8i8_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v10, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16(
+  <vscale x 8 x i8>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv8i8_nxv8bf16(<vscale x 8 x i8> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv8i8_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv8i8.nxv8bf16(
+    <vscale x 8 x i8> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_xu.f.w_nxv16i8_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v12, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16(
+  <vscale x 16 x i8>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv16i8_nxv16bf16(<vscale x 16 x i8> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv16i8_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv16i8.nxv16bf16(
+    <vscale x 16 x i8> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_xu.f.w_nxv32i8_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_xu.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vfncvt.xu.f.w v16, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 32 x i8> %a
+}
+
+declare <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16(
+  <vscale x 32 x i8>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x i8> @intrinsic_vfncvt_mask_xu.f.w_nxv32i8_nxv32bf16(<vscale x 32 x i8> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfncvt_mask_xu.f.w_nxv32i8_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfncvt.xu.f.w v8, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x i8> @llvm.riscv.vfncvt.xu.f.w.mask.nxv32i8.nxv32bf16(
+    <vscale x 32 x i8> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x i8> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfnmacc-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfnmacc-bf.ll
new file mode 100644
index 0000000000000..4bf643003a38c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfnmacc-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmacc_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmacc_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmacc_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmacc_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmacc_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmacc_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmacc_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmacc.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmacc_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmacc.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmacc_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmacc.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmacc_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmacc.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmacc_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfnmacc_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmacc.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmacc_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfnmacc_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmacc.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmacc_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfnmacc_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmacc.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmacc_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfnmacc_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmacc.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmacc_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfnmacc_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmacc_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmacc.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmacc.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfnmadd-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfnmadd-bf.ll
new file mode 100644
index 0000000000000..7dcaa1c24e6de
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfnmadd-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmadd_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmadd_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmadd_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmadd_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmadd.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmadd.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmadd_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmadd.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmadd.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmadd_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfnmadd_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmadd.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmadd_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfnmadd_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmadd.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmadd_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfnmadd_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmadd.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmadd_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfnmadd_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmadd.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmadd_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfnmadd_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmadd_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmadd.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmadd.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfnmsac-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfnmsac-bf.ll
new file mode 100644
index 0000000000000..9528f80453398
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfnmsac-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsac_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsac_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsac_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsac_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsac_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsac_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsac_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmsac.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsac_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmsac.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsac_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmsac.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsac_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmsac.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsac_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfnmsac_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsac.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsac_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfnmsac_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsac.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsac_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfnmsac_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsac.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsac_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfnmsac_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsac.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsac_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfnmsac_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsac_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmsac.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsac.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfnmsub-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfnmsub-bf.ll
new file mode 100644
index 0000000000000..dcbb9ced92db7
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfnmsub-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsub_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsub_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsub_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsub_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmsub.vv v8, v10, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmsub.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsub_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmsub.vv v8, v12, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmsub.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat>  @intrinsic_vfnmsub_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfnmsub_mask_vf_nxv1bf16_bf16_nxv1bf16(<vscale x 1 x bfloat> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vf_nxv1bf16_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfnmsub.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat>  @intrinsic_vfnmsub_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfnmsub_mask_vf_nxv2bf16_bf16_nxv2bf16(<vscale x 2 x bfloat> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vf_nxv2bf16_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfnmsub.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat>  @intrinsic_vfnmsub_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfnmsub_mask_vf_nxv4bf16_bf16_nxv4bf16(<vscale x 4 x bfloat> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vf_nxv4bf16_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfnmsub.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat>  @intrinsic_vfnmsub_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfnmsub_mask_vf_nxv8bf16_bf16_nxv8bf16(<vscale x 8 x bfloat> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vf_nxv8bf16_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfnmsub.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat>  @intrinsic_vfnmsub_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfnmsub_mask_vf_nxv16bf16_bf16_nxv16bf16(<vscale x 16 x bfloat> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfnmsub_mask_vf_nxv16bf16_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfnmsub.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfnmsub.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0);
+
+  ret <vscale x 16 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfrec7-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfrec7-bf.ll
new file mode 100644
index 0000000000000..1211415ffe432
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfrec7-bf.ll
@@ -0,0 +1,282 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrec7_v_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrec7_mask_v_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrec7.mask.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrec7_v_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrec7_mask_v_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrec7.mask.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrec7_v_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrec7_mask_v_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrec7.mask.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrec7_v_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrec7_mask_v_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrec7.mask.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrec7_v_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrec7_mask_v_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrec7.mask.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrec7_v_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_v_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfrec7.v v8, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen 0, iXLen %1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrec7_mask_v_nxv32bf16_nxv32bf16(<vscale x 32 x i1> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrec7_mask_v_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfrec7.v v8, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrec7.mask.nxv32bf16(
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %0,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfrsqrt7-bf16.ll b/llvm/test/CodeGen/RISCV/rvv/vfrsqrt7-bf16.ll
new file mode 100644
index 0000000000000..4626b865ab454
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfrsqrt7-bf16.ll
@@ -0,0 +1,264 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrsqrt7_v_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrsqrt7_v_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrsqrt7_v_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrsqrt7_v_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrsqrt7_v_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrsqrt7_v_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_v_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfrsqrt7.v v8, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrsqrt7_mask_v_nxv32bf16_nxv32bf16(<vscale x 32 x i1> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfrsqrt7_mask_v_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfrsqrt7.v v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrsqrt7.mask.nxv32bf16(
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %0,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfrsub-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfrsub-bf.ll
new file mode 100644
index 0000000000000..54a6d48cfcc5b
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfrsub-bf.ll
@@ -0,0 +1,282 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrsub_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfrsub_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfrsub.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrsub_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfrsub_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfrsub.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrsub_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfrsub_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfrsub.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrsub_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfrsub_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfrsub.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrsub_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfrsub_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfrsub.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrsub_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfrsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfrsub_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfrsub_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfrsub.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfrsub.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsgnj-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfsgnj-bf.ll
new file mode 100644
index 0000000000000..2cd698d9aaa3c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsgnj-bf.ll
@@ -0,0 +1,571 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnj_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnj_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnj_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnj_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnj_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnj_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnj.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnj_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnj_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnj.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnj_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnj.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnj_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnj.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnj_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnj.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnj_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnj.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnj_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnj.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnj_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnj_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnj.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnj.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsgnjn-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfsgnjn-bf.ll
new file mode 100644
index 0000000000000..08340becc9ed4
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsgnjn-bf.ll
@@ -0,0 +1,571 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjn_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjn_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjn_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjn_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjn_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjn_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnjn.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjn_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnjn.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjn_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjn_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjn_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjn_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjn_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjn_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnjn.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjn_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjn_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnjn.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjn.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsgnjx-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfsgnjx-bf.ll
new file mode 100644
index 0000000000000..e51a42e2b8cea
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsgnjx-bf.ll
@@ -0,0 +1,571 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjx_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjx_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjx_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjx_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjx_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjx_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnjx.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjx_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnjx.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjx_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjx_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjx_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjx_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjx_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjx_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsgnjx.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsgnjx_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsgnjx_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsgnjx.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsgnjx.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfslide1down-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfslide1down-bf.ll
new file mode 100644
index 0000000000000..c65719c3a4c1a
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfslide1down-bf.ll
@@ -0,0 +1,288 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfslide1down_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1down.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfslide1down_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1down.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfslide1down_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1down.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfslide1down_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1down.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfslide1down_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1down.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfslide1down_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfslide1down.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfslide1down_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1down_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfslide1down.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1down.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfslide1up-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfslide1up-bf.ll
new file mode 100644
index 0000000000000..57a48986fdfcd
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfslide1up-bf.ll
@@ -0,0 +1,294 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfslide1up_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v9, v8, fa0
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfslide1up.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfslide1up_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v9, v8, fa0
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfslide1up.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfslide1up_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v9, v8, fa0
+; CHECK-NEXT:    vmv.v.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfslide1up.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfslide1up_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v10, v8, fa0
+; CHECK-NEXT:    vmv.v.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfslide1up.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfslide1up_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v12, v8, fa0
+; CHECK-NEXT:    vmv.v.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfslide1up.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfslide1up_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfslide1up.vf v16, v8, fa0
+; CHECK-NEXT:    vmv.v.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfslide1up_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfslide1up_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfslide1up.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfslide1up.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfsub-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfsub-bf.ll
new file mode 100644
index 0000000000000..aea75211b70b5
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfsub-bf.ll
@@ -0,0 +1,559 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsub_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv1bf16_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsub_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv2bf16_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsub_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv4bf16_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsub_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv8bf16_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v10, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsub_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv16bf16_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v12, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsub_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsub.vv v8, v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsub_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, <vscale x 32 x bfloat> %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vv_nxv32bf16_nxv32bf16_nxv32bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl8re16.v v24, (a0)
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsub.vv v8, v16, v24, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.nxv32bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    <vscale x 32 x bfloat> %2,
+    <vscale x 32 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsub_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfsub_mask_vf_nxv1bf16_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv1bf16_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfsub.mask.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsub_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfsub_mask_vf_nxv2bf16_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv2bf16_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfsub.mask.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsub_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfsub_mask_vf_nxv4bf16_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv4bf16_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfsub.mask.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsub_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfsub_mask_vf_nxv8bf16_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv8bf16_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfsub.mask.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsub_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfsub_mask_vf_nxv16bf16_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv16bf16_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfsub.mask.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsub_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, ma
+; CHECK-NEXT:    vfsub.vf v8, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x bfloat> %0,
+    bfloat %1,
+    iXLen 7, iXLen %2)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x bfloat>,
+  bfloat,
+  <vscale x 32 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfsub_mask_vf_nxv32bf16_nxv32bf16_bf16(<vscale x 32 x bfloat> %0, <vscale x 32 x bfloat> %1, bfloat %2, <vscale x 32 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfsub_mask_vf_nxv32bf16_nxv32bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m8, ta, mu
+; CHECK-NEXT:    vfsub.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfsub.mask.nxv32bf16.bf16(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x bfloat> %1,
+    bfloat %2,
+    <vscale x 32 x i1> %3,
+    iXLen 7, iXLen %4, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwadd-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwadd-bf.ll
new file mode 100644
index 0000000000000..62feac824efad
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwadd-bf.ll
@@ -0,0 +1,519 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwadd.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwadd.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v9
+; CHECK-NEXT:    vmv1r.v v11, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vv v8, v11, v10
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v10
+; CHECK-NEXT:    vmv2r.v v14, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vv v8, v14, v12
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v12
+; CHECK-NEXT:    vmv4r.v v20, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vv v8, v20, v16
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwadd.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd_mask_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.mask.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwadd.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd_mask_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.mask.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vf v8, v10, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd_mask_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.mask.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vf v8, v12, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd_mask_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.mask.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwadd.vf v8, v16, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd_mask_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd_mask_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.mask.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwadd-w-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwadd-w-bf.ll
new file mode 100644
index 0000000000000..69c9a4ea75c97
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwadd-w-bf.ll
@@ -0,0 +1,773 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_wv_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwadd.wv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_mask_wv_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x float> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_wv_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwadd.wv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_mask_wv_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x float> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_wv_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwadd.wv v8, v8, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_mask_wv_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x float> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_wv_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwadd.wv v8, v8, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_mask_wv_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x float> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_wv_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfwadd.wv v8, v8, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_mask_wv_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x float> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl4re16.v v24, (a0)
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v16, v24, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_wf_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wf_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_mask_wf_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, <vscale x 1 x float> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_wf_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wf_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_mask_wf_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, <vscale x 2 x float> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_wf_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wf_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_mask_wf_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, <vscale x 4 x float> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_wf_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wf_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_mask_wf_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, <vscale x 8 x float> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_wf_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wf_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.nxv16f32.bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_mask_wf_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, <vscale x 16 x float> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_mask_wv_tie_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_tie_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_mask_wv_tie_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_tie_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_mask_wv_tie_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_tie_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v8, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_mask_wv_tie_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_tie_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v8, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_mask_wv_tie_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wv_tie_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.wv v8, v8, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_mask_wf_tie_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_tie_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_mask_wf_tie_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_tie_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_mask_wf_tie_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_tie_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_mask_wf_tie_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_tie_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+define <vscale x 16 x float> @intrinsic_vfwadd.w_mask_wf_tie_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_mask_wf_tie_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwadd.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwadd.w.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwadd.w_wv_untie_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_untie_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwadd.wv v9, v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwadd.w.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %1,
+    <vscale x 1 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwadd.w_wv_untie_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_untie_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwadd.wv v9, v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwadd.w.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %1,
+    <vscale x 2 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwadd.w_wv_untie_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_untie_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwadd.wv v10, v10, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv2r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwadd.w.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %1,
+    <vscale x 4 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwadd.w_wv_untie_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwadd.w_wv_untie_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwadd.wv v12, v12, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv4r.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwadd.w.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %1,
+    <vscale x 8 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-x.ll b/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-x.ll
new file mode 100644
index 0000000000000..b7df45bad36e6
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-x.ll
@@ -0,0 +1,264 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i8>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv1bf16_nxv1i8(<vscale x 1 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv1bf16_nxv1i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfwcvt.f.x.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv1bf16.nxv1i8(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i8>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv1bf16_nxv1i8(<vscale x 1 x bfloat> %0, <vscale x 1 x i8> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv1bf16_nxv1i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv1bf16.nxv1i8(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x i8> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i8>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv2bf16_nxv2i8(<vscale x 2 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv2bf16_nxv2i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfwcvt.f.x.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv2bf16.nxv2i8(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i8>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv2bf16_nxv2i8(<vscale x 2 x bfloat> %0, <vscale x 2 x i8> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv2bf16_nxv2i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv2bf16.nxv2i8(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x i8> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i8>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv4bf16_nxv4i8(<vscale x 4 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv4bf16_nxv4i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfwcvt.f.x.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv4bf16.nxv4i8(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i8>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv4bf16_nxv4i8(<vscale x 4 x bfloat> %0, <vscale x 4 x i8> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv4bf16_nxv4i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv4bf16.nxv4i8(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x i8> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i8>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv8bf16_nxv8i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv8bf16_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v8
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv8bf16.nxv8i8(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv8bf16_nxv8i8(<vscale x 8 x bfloat> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv8bf16_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv8bf16.nxv8i8(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x i8> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i8>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv16bf16_nxv16i8(<vscale x 16 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv16bf16_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v8
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv16bf16.nxv16i8(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv16bf16_nxv16i8(<vscale x 16 x bfloat> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv16bf16_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv16bf16.nxv16i8(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x i8> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i8>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfwcvt_f.x.v_nxv32bf16_nxv32i8(<vscale x 32 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.x.v_nxv32bf16_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v8
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.nxv32bf16.nxv32i8(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfwcvt_mask_f.x.v_nxv32bf16_nxv32i8(<vscale x 32 x bfloat> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.x.v_nxv32bf16_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.x.v v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.x.v.mask.nxv32bf16.nxv32i8(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x i8> %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-xu.ll b/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-xu.ll
new file mode 100644
index 0000000000000..c370261a77bc0
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwcvt-bf-xu.ll
@@ -0,0 +1,264 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i8>,
+  iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv1bf16_nxv1i8(<vscale x 1 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv1bf16_nxv1i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, ma
+; CHECK-NEXT:    vfwcvt.f.xu.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv1bf16.nxv1i8(
+    <vscale x 1 x bfloat> poison,
+    <vscale x 1 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i8>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 1 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv1bf16_nxv1i8(<vscale x 1 x bfloat> %0, <vscale x 1 x i8> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv1bf16_nxv1i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf8, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv1bf16.nxv1i8(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x i8> %1,
+    <vscale x 1 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i8>,
+  iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv2bf16_nxv2i8(<vscale x 2 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv2bf16_nxv2i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, ma
+; CHECK-NEXT:    vfwcvt.f.xu.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv2bf16.nxv2i8(
+    <vscale x 2 x bfloat> poison,
+    <vscale x 2 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i8>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 2 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv2bf16_nxv2i8(<vscale x 2 x bfloat> %0, <vscale x 2 x i8> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv2bf16_nxv2i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv2bf16.nxv2i8(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x i8> %1,
+    <vscale x 2 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i8>,
+  iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv4bf16_nxv4i8(<vscale x 4 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv4bf16_nxv4i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, ma
+; CHECK-NEXT:    vfwcvt.f.xu.v v9, v8
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv4bf16.nxv4i8(
+    <vscale x 4 x bfloat> poison,
+    <vscale x 4 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i8>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 4 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv4bf16_nxv4i8(<vscale x 4 x bfloat> %0, <vscale x 4 x i8> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv4bf16_nxv4i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, mf2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v9, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv4bf16.nxv4i8(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x i8> %1,
+    <vscale x 4 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i8>,
+  iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv8bf16_nxv8i8(<vscale x 8 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv8bf16_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v8
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv8bf16.nxv8i8(
+    <vscale x 8 x bfloat> poison,
+    <vscale x 8 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i8>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 8 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv8bf16_nxv8i8(<vscale x 8 x bfloat> %0, <vscale x 8 x i8> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv8bf16_nxv8i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m1, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v10, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv8bf16.nxv8i8(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x i8> %1,
+    <vscale x 8 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i8>,
+  iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv16bf16_nxv16i8(<vscale x 16 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv16bf16_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v8
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv16bf16.nxv16i8(
+    <vscale x 16 x bfloat> poison,
+    <vscale x 16 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i8>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 16 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv16bf16_nxv16i8(<vscale x 16 x bfloat> %0, <vscale x 16 x i8> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv16bf16_nxv16i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m2, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v12, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv16bf16.nxv16i8(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x i8> %1,
+    <vscale x 16 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i8>,
+  iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfwcvt_f.xu.v_nxv32bf16_nxv32i8(<vscale x 32 x i8> %0, iXLen %1) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_f.xu.v_nxv32bf16_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v8
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.nxv32bf16.nxv32i8(
+    <vscale x 32 x bfloat> poison,
+    <vscale x 32 x i8> %0,
+    iXLen %1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
+declare <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8(
+  <vscale x 32 x bfloat>,
+  <vscale x 32 x i8>,
+  <vscale x 32 x i1>,
+  iXLen, iXLen);
+
+define <vscale x 32 x bfloat> @intrinsic_vfwcvt_mask_f.xu.v_nxv32bf16_nxv32i8(<vscale x 32 x bfloat> %0, <vscale x 32 x i8> %1, <vscale x 32 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwcvt_mask_f.xu.v_nxv32bf16_nxv32i8:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e8alt, m4, ta, mu
+; CHECK-NEXT:    vfwcvt.f.xu.v v8, v16, v0.t
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x bfloat> @llvm.riscv.vfwcvt.f.xu.v.mask.nxv32bf16.nxv32i8(
+    <vscale x 32 x bfloat> %0,
+    <vscale x 32 x i8> %1,
+    <vscale x 32 x i1> %2,
+    iXLen %3, iXLen 1)
+
+  ret <vscale x 32 x bfloat> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwmsac-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwmsac-bf.ll
new file mode 100644
index 0000000000000..a3f667818ab0a
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwmsac-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwmsac_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwmsac_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwmsac_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwmsac_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwmsac_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwmsac.vv v8, v10, v11
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwmsac_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwmsac.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwmsac_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwmsac.vv v8, v12, v14
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwmsac_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwmsac.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwmsac_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwmsac.vv v8, v16, v20
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwmsac_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwmsac.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwmsac_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwmsac_mask_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmsac.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwmsac_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwmsac_mask_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmsac.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwmsac_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwmsac_mask_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmsac.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwmsac_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwmsac_mask_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmsac.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwmsac_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwmsac_mask_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmsac_mask_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwmsac.vf v8, fa0, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmsac.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwmul-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwmul-bf.ll
new file mode 100644
index 0000000000000..577b93af7a918
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwmul-bf.ll
@@ -0,0 +1,519 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwmul_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwmul.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwmul_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwmul.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwmul_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwmul.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwmul_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwmul.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwmul_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v9
+; CHECK-NEXT:    vmv1r.v v11, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vv v8, v11, v10
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwmul_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwmul.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwmul_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v10
+; CHECK-NEXT:    vmv2r.v v14, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vv v8, v14, v12
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwmul_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwmul.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwmul_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v12
+; CHECK-NEXT:    vmv4r.v v20, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vv v8, v20, v16
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwmul_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwmul.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwmul_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwmul.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmul.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwmul_mask_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwmul.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwmul.mask.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwmul_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwmul.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmul.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwmul_mask_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwmul.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwmul.mask.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwmul_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vf v8, v10, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmul.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwmul_mask_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwmul.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwmul.mask.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwmul_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vf v8, v12, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmul.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwmul_mask_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwmul.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwmul.mask.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwmul_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwmul.vf v8, v16, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmul.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwmul_mask_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwmul_mask_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwmul.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwmul.mask.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwnmacc-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwnmacc-bf.ll
new file mode 100644
index 0000000000000..1e05e4c7acf25
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwnmacc-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmacc_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwnmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmacc_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwnmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmacc_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwnmacc.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmacc_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwnmacc.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmacc_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwnmacc.vv v8, v10, v11
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmacc_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwnmacc.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmacc_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwnmacc.vv v8, v12, v14
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmacc_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwnmacc.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmacc_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwnmacc.vv v8, v16, v20
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmacc_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwnmacc.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmacc_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwnmacc_mask_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmacc_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwnmacc_mask_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmacc_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwnmacc_mask_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmacc_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwnmacc_mask_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmacc_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwnmacc_mask_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmacc_mask_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwnmacc.vf v8, fa0, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwnmsac-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwnmsac-bf.ll
new file mode 100644
index 0000000000000..223ad4f7483f6
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwnmsac-bf.ll
@@ -0,0 +1,506 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmsac_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwnmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmsac_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwnmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmsac_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwnmsac.vv v8, v9, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmsac_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwnmsac.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmsac_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwnmsac.vv v8, v10, v11
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmsac_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwnmsac.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmsac_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwnmsac.vv v8, v12, v14
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmsac_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwnmsac.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmsac_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwnmsac.vv v8, v16, v20
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmsac_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwnmsac.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float>  @intrinsic_vfwnmsac_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, ma
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwnmsac_mask_vf_nxv1f32_bf16_nxv1bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vf_nxv1f32_bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, tu, mu
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float>  @intrinsic_vfwnmsac_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, ma
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwnmsac_mask_vf_nxv2f32_bf16_nxv2bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vf_nxv2f32_bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, tu, mu
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float>  @intrinsic_vfwnmsac_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, ma
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwnmsac_mask_vf_nxv4f32_bf16_nxv4bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vf_nxv4f32_bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, tu, mu
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float>  @intrinsic_vfwnmsac_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, ma
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwnmsac_mask_vf_nxv8f32_bf16_nxv8bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vf_nxv8f32_bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, tu, mu
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float>  @intrinsic_vfwnmsac_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, ma
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen 0, iXLen %3, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwnmsac_mask_vf_nxv16f32_bf16_nxv16bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwnmsac_mask_vf_nxv16f32_bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, tu, mu
+; CHECK-NEXT:    vfwnmsac.vf v8, fa0, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 0)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwsub-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwsub-bf.ll
new file mode 100644
index 0000000000000..d993e4e610d2c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwsub-bf.ll
@@ -0,0 +1,519 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwsub.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub_mask_vv_nxv1f32_nxv1bf16_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vv_nxv1f32_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwsub.vv v10, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub_mask_vv_nxv2f32_nxv2bf16_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vv_nxv2f32_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.vv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v9
+; CHECK-NEXT:    vmv1r.v v11, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vv v8, v11, v10
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub_mask_vv_nxv4f32_nxv4bf16_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vv_nxv4f32_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.vv v8, v10, v11, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v10
+; CHECK-NEXT:    vmv2r.v v14, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vv v8, v14, v12
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub_mask_vv_nxv8f32_nxv8bf16_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vv_nxv8f32_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.vv v8, v12, v14, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v12
+; CHECK-NEXT:    vmv4r.v v20, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vv v8, v20, v16
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub_mask_vv_nxv16f32_nxv16bf16_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vv_nxv16f32_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.vv v8, v16, v20, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwsub.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub_mask_vf_nxv1f32_nxv1bf16_bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vf_nxv1f32_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.mask.nxv1f32.nxv1bf16.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwsub.vf v9, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub_mask_vf_nxv2f32_nxv2bf16_bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vf_nxv2f32_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.vf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.mask.nxv2f32.nxv2bf16.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmv1r.v v10, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vf v8, v10, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub_mask_vf_nxv4f32_nxv4bf16_bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vf_nxv4f32_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.vf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.mask.nxv4f32.nxv4bf16.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmv2r.v v12, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vf v8, v12, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub_mask_vf_nxv8f32_nxv8bf16_bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vf_nxv8f32_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.vf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.mask.nxv8f32.nxv8bf16.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmv4r.v v16, v8
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vfwsub.vf v8, v16, fa0
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub_mask_vf_nxv16f32_nxv16bf16_bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub_mask_vf_nxv16f32_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.vf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.mask.nxv16f32.nxv16bf16.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vfwsub-w-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vfwsub-w-bf.ll
new file mode 100644
index 0000000000000..11066d9487684
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vfwsub-w-bf.ll
@@ -0,0 +1,773 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_wv_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwsub.wv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_mask_wv_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x float> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %1,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_wv_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwsub.wv v8, v8, v9
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_mask_wv_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x float> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v9, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %1,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_wv_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwsub.wv v8, v8, v10
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_mask_wv_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x float> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v10, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %1,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_wv_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwsub.wv v8, v8, v12
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_mask_wv_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x float> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v12, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %1,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_wv_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfwsub.wv v8, v8, v16
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_mask_wv_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x float> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vl4re16.v v24, (a0)
+; CHECK-NEXT:    fsrmi a0, 0
+; CHECK-NEXT:    vsetvli zero, a1, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v16, v24, v0.t
+; CHECK-NEXT:    fsrm a0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %1,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_wf_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wf_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16(
+  <vscale x 1 x float>,
+  <vscale x 1 x float>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_mask_wf_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, <vscale x 1 x float> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_wf_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wf_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16(
+  <vscale x 2 x float>,
+  <vscale x 2 x float>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_mask_wf_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, <vscale x 2 x float> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v9, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_wf_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wf_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16(
+  <vscale x 4 x float>,
+  <vscale x 4 x float>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_mask_wf_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, <vscale x 4 x float> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v10, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_wf_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wf_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16(
+  <vscale x 8 x float>,
+  <vscale x 8 x float>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_mask_wf_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, <vscale x 8 x float> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v12, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  bfloat,
+  iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_wf_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wf_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.nxv16f32.bf16(
+    <vscale x 16 x float> poison,
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16(
+  <vscale x 16 x float>,
+  <vscale x 16 x float>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen, iXLen, iXLen);
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_mask_wf_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, <vscale x 16 x float> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v16, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen 0, iXLen %4, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_mask_wv_tie_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x float> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_tie_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %0,
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_mask_wv_tie_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x float> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_tie_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v8, v9, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %0,
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_mask_wv_tie_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x float> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_tie_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v8, v10, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %0,
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_mask_wv_tie_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x float> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_tie_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v8, v12, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %0,
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_mask_wv_tie_nxv16f32_nxv16f32_nxv16bf16(<vscale x 16 x float> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wv_tie_nxv16f32_nxv16f32_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.wv v8, v8, v16, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.nxv16bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %0,
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_mask_wf_tie_nxv1f32_nxv1f32_bf16(<vscale x 1 x float> %0, bfloat %1, <vscale x 1 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_tie_nxv1f32_nxv1f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.mask.nxv1f32.bf16(
+    <vscale x 1 x float> %0,
+    <vscale x 1 x float> %0,
+    bfloat %1,
+    <vscale x 1 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_mask_wf_tie_nxv2f32_nxv2f32_bf16(<vscale x 2 x float> %0, bfloat %1, <vscale x 2 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_tie_nxv2f32_nxv2f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.mask.nxv2f32.bf16(
+    <vscale x 2 x float> %0,
+    <vscale x 2 x float> %0,
+    bfloat %1,
+    <vscale x 2 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_mask_wf_tie_nxv4f32_nxv4f32_bf16(<vscale x 4 x float> %0, bfloat %1, <vscale x 4 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_tie_nxv4f32_nxv4f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.mask.nxv4f32.bf16(
+    <vscale x 4 x float> %0,
+    <vscale x 4 x float> %0,
+    bfloat %1,
+    <vscale x 4 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_mask_wf_tie_nxv8f32_nxv8f32_bf16(<vscale x 8 x float> %0, bfloat %1, <vscale x 8 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_tie_nxv8f32_nxv8f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.mask.nxv8f32.bf16(
+    <vscale x 8 x float> %0,
+    <vscale x 8 x float> %0,
+    bfloat %1,
+    <vscale x 8 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 8 x float> %a
+}
+
+define <vscale x 16 x float> @intrinsic_vfwsub.w_mask_wf_tie_nxv16f32_nxv16f32_bf16(<vscale x 16 x float> %0, bfloat %1, <vscale x 16 x i1> %2, iXLen %3) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_mask_wf_tie_nxv16f32_nxv16f32_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vfwsub.wf v8, v8, fa0, v0.t
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfwsub.w.mask.nxv16f32.bf16(
+    <vscale x 16 x float> %0,
+    <vscale x 16 x float> %0,
+    bfloat %1,
+    <vscale x 16 x i1> %2,
+    iXLen 0, iXLen %3, iXLen 1)
+
+  ret <vscale x 16 x float> %a
+}
+
+define <vscale x 1 x float> @intrinsic_vfwsub.w_wv_untie_nxv1f32_nxv1f32_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_untie_nxv1f32_nxv1f32_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vfwsub.wv v9, v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfwsub.w.nxv1f32.nxv1bf16(
+    <vscale x 1 x float> poison,
+    <vscale x 1 x float> %1,
+    <vscale x 1 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 1 x float> %a
+}
+
+define <vscale x 2 x float> @intrinsic_vfwsub.w_wv_untie_nxv2f32_nxv2f32_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_untie_nxv2f32_nxv2f32_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vfwsub.wv v9, v9, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv1r.v v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfwsub.w.nxv2f32.nxv2bf16(
+    <vscale x 2 x float> poison,
+    <vscale x 2 x float> %1,
+    <vscale x 2 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 2 x float> %a
+}
+
+define <vscale x 4 x float> @intrinsic_vfwsub.w_wv_untie_nxv4f32_nxv4f32_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_untie_nxv4f32_nxv4f32_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vfwsub.wv v10, v10, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv2r.v v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfwsub.w.nxv4f32.nxv4bf16(
+    <vscale x 4 x float> poison,
+    <vscale x 4 x float> %1,
+    <vscale x 4 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 4 x float> %a
+}
+
+define <vscale x 8 x float> @intrinsic_vfwsub.w_wv_untie_nxv8f32_nxv8f32_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x float> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vfwsub.w_wv_untie_nxv8f32_nxv8f32_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    fsrmi a1, 0
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vfwsub.wv v12, v12, v8
+; CHECK-NEXT:    fsrm a1
+; CHECK-NEXT:    vmv4r.v v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfwsub.w.nxv8f32.nxv8bf16(
+    <vscale x 8 x float> poison,
+    <vscale x 8 x float> %1,
+    <vscale x 8 x bfloat> %0,
+    iXLen 0, iXLen %2)
+
+  ret <vscale x 8 x float> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmfeq-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmfeq-bf.ll
new file mode 100644
index 0000000000000..9bd859b3452f2
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmfeq-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfeq_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfeq_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    vmfeq.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfeq_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfeq_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    vmfeq.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfeq_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfeq_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfeq.vv v0, v8, v9
+; CHECK-NEXT:    vmfeq.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfeq_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfeq.vv v0, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfeq_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmfeq.vv v0, v8, v10
+; CHECK-NEXT:    vmfeq.vv v14, v10, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfeq_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfeq.vv v0, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfeq_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmfeq.vv v0, v8, v12
+; CHECK-NEXT:    vmfeq.vv v20, v12, v16, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfeq_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfeq.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfeq.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfeq_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfeq.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfeq.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfeq_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfeq.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfeq.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfeq_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfeq.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfeq.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfeq_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfeq.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfeq.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfeq_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfeq.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfeq.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfeq_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfeq.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfeq.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfeq_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmfeq.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfeq.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfeq_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfeq.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfeq.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfeq_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfeq_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmfeq.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfeq.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmfge-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmfge-bf.ll
new file mode 100644
index 0000000000000..73946dc1a744c
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmfge-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfge_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfge_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    vmfle.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfge_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfge_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    vmfle.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfge_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfge_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v9, v8
+; CHECK-NEXT:    vmfle.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfge_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v10, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfge_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmfle.vv v0, v10, v8
+; CHECK-NEXT:    vmfle.vv v14, v12, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfge_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v12, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfge_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmfle.vv v0, v12, v8
+; CHECK-NEXT:    vmfle.vv v20, v16, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfge_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfge.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfge.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfge_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfge.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfge.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfge_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfge.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfge.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfge_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfge.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfge.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfge_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfge.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfge.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfge_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfge.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfge.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfge_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfge.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfge.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfge_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmfge.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfge.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfge_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfge.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfge.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfge_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfge_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmfge.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfge.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmfgt-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmfgt-bf.ll
new file mode 100644
index 0000000000000..fac324ca5c125
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmfgt-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfgt_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfgt_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    vmflt.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfgt_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfgt_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    vmflt.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfgt_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfgt_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v9, v8
+; CHECK-NEXT:    vmflt.vv v11, v10, v9, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfgt_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v10, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfgt_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmflt.vv v0, v10, v8
+; CHECK-NEXT:    vmflt.vv v14, v12, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfgt_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v12, v8
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfgt_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmflt.vv v0, v12, v8
+; CHECK-NEXT:    vmflt.vv v20, v16, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfgt_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfgt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfgt.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfgt_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfgt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfgt.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfgt_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfgt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfgt.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfgt_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfgt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfgt.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfgt_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfgt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfgt.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfgt_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfgt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfgt.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfgt_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfgt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfgt.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfgt_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmfgt.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfgt.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfgt_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfgt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfgt.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfgt_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfgt_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmfgt.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfgt.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmfle-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmfle-bf.ll
new file mode 100644
index 0000000000000..8356b7bbd3ff7
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmfle-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfle_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfle_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    vmfle.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfle_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfle_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    vmfle.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfle_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfle_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfle.vv v0, v8, v9
+; CHECK-NEXT:    vmfle.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfle_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfle_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmfle.vv v0, v8, v10
+; CHECK-NEXT:    vmfle.vv v14, v10, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfle_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfle.vv v0, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfle_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmfle.vv v0, v8, v12
+; CHECK-NEXT:    vmfle.vv v20, v12, v16, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfle_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfle.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfle.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfle_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfle.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfle.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfle_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfle.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfle.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfle_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfle.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfle.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfle_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfle.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfle.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfle_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfle.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfle.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfle_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfle.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfle.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfle_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmfle.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfle.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfle_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfle.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfle.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfle_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfle_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmfle.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfle.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmflt-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmflt-bf.ll
new file mode 100644
index 0000000000000..2e1bcc5e87bfc
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmflt-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmflt_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmflt_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    vmflt.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmflt_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmflt_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    vmflt.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmflt_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmflt_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmflt.vv v0, v8, v9
+; CHECK-NEXT:    vmflt.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmflt_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmflt_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmflt.vv v0, v8, v10
+; CHECK-NEXT:    vmflt.vv v14, v10, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmflt_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmflt.vv v0, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmflt_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmflt.vv v0, v8, v12
+; CHECK-NEXT:    vmflt.vv v20, v12, v16, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmflt_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmflt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmflt.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmflt_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmflt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmflt.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmflt_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmflt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmflt.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmflt_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmflt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmflt.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmflt_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmflt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmflt.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmflt_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmflt.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmflt.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmflt_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmflt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmflt.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmflt_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmflt.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmflt.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmflt_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmflt.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmflt.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmflt_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmflt_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmflt.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmflt.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
diff --git a/llvm/test/CodeGen/RISCV/rvv/vmfne-bf.ll b/llvm/test/CodeGen/RISCV/rvv/vmfne-bf.ll
new file mode 100644
index 0000000000000..283ffc500fdde
--- /dev/null
+++ b/llvm/test/CodeGen/RISCV/rvv/vmfne-bf.ll
@@ -0,0 +1,496 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: sed 's/iXLen/i32/g' %s | llc -mtriple=riscv32 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=ilp32d | FileCheck %s
+; RUN: sed 's/iXLen/i64/g' %s | llc -mtriple=riscv64 -mattr=+v,+experimental-zvfbfa \
+; RUN:   -verify-machineinstrs -target-abi=lp64d | FileCheck %s
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16(
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfne_vv_nxv1bf16_nxv1bf16(<vscale x 1 x bfloat> %0, <vscale x 1 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16(
+    <vscale x 1 x bfloat> %0,
+    <vscale x 1 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x bfloat>,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfne_mask_vv_nxv1bf16_nxv1bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, <vscale x 1 x bfloat> %2, <vscale x 1 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vv_nxv1bf16_nxv1bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    vmfne.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16(
+    <vscale x 1 x bfloat> %1,
+    <vscale x 1 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %2,
+    <vscale x 1 x bfloat> %3,
+    <vscale x 1 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16(
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfne_vv_nxv2bf16_nxv2bf16(<vscale x 2 x bfloat> %0, <vscale x 2 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16(
+    <vscale x 2 x bfloat> %0,
+    <vscale x 2 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x bfloat>,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfne_mask_vv_nxv2bf16_nxv2bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, <vscale x 2 x bfloat> %2, <vscale x 2 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vv_nxv2bf16_nxv2bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    vmfne.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16(
+    <vscale x 2 x bfloat> %1,
+    <vscale x 2 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %2,
+    <vscale x 2 x bfloat> %3,
+    <vscale x 2 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16(
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfne_vv_nxv4bf16_nxv4bf16(<vscale x 4 x bfloat> %0, <vscale x 4 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16(
+    <vscale x 4 x bfloat> %0,
+    <vscale x 4 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x bfloat>,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfne_mask_vv_nxv4bf16_nxv4bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, <vscale x 4 x bfloat> %2, <vscale x 4 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vv_nxv4bf16_nxv4bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmfne.vv v0, v8, v9
+; CHECK-NEXT:    vmfne.vv v11, v9, v10, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16(
+    <vscale x 4 x bfloat> %1,
+    <vscale x 4 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %2,
+    <vscale x 4 x bfloat> %3,
+    <vscale x 4 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16(
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfne_vv_nxv8bf16_nxv8bf16(<vscale x 8 x bfloat> %0, <vscale x 8 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfne.vv v0, v8, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16(
+    <vscale x 8 x bfloat> %0,
+    <vscale x 8 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x bfloat>,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfne_mask_vv_nxv8bf16_nxv8bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, <vscale x 8 x bfloat> %2, <vscale x 8 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vv_nxv8bf16_nxv8bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v14, v0
+; CHECK-NEXT:    vmfne.vv v0, v8, v10
+; CHECK-NEXT:    vmfne.vv v14, v10, v12, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v14
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16(
+    <vscale x 8 x bfloat> %1,
+    <vscale x 8 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %2,
+    <vscale x 8 x bfloat> %3,
+    <vscale x 8 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16(
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfne_vv_nxv16bf16_nxv16bf16(<vscale x 16 x bfloat> %0, <vscale x 16 x bfloat> %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfne.vv v0, v8, v12
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16(
+    <vscale x 16 x bfloat> %0,
+    <vscale x 16 x bfloat> %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x bfloat>,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfne_mask_vv_nxv16bf16_nxv16bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, <vscale x 16 x bfloat> %2, <vscale x 16 x bfloat> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vv_nxv16bf16_nxv16bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v20, v0
+; CHECK-NEXT:    vmfne.vv v0, v8, v12
+; CHECK-NEXT:    vmfne.vv v20, v12, v16, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v20
+; CHECK-NEXT:    ret
+entry:
+  %mask = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16(
+    <vscale x 16 x bfloat> %1,
+    <vscale x 16 x bfloat> %2,
+    iXLen %4)
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %2,
+    <vscale x 16 x bfloat> %3,
+    <vscale x 16 x i1> %mask,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.bf16(
+  <vscale x 1 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfne_vf_nxv1bf16_bf16(<vscale x 1 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, ma
+; CHECK-NEXT:    vmfne.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfne.nxv1bf16.bf16(
+    <vscale x 1 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16(
+  <vscale x 1 x i1>,
+  <vscale x 1 x bfloat>,
+  bfloat,
+  <vscale x 1 x i1>,
+  iXLen);
+
+define <vscale x 1 x i1> @intrinsic_vmfne_mask_vf_nxv1bf16_bf16(<vscale x 1 x i1> %0, <vscale x 1 x bfloat> %1, bfloat %2, <vscale x 1 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vf_nxv1bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf4, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfne.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x i1> @llvm.riscv.vmfne.mask.nxv1bf16.bf16(
+    <vscale x 1 x i1> %0,
+    <vscale x 1 x bfloat> %1,
+    bfloat %2,
+    <vscale x 1 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 1 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.bf16(
+  <vscale x 2 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfne_vf_nxv2bf16_bf16(<vscale x 2 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, ma
+; CHECK-NEXT:    vmfne.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfne.nxv2bf16.bf16(
+    <vscale x 2 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16(
+  <vscale x 2 x i1>,
+  <vscale x 2 x bfloat>,
+  bfloat,
+  <vscale x 2 x i1>,
+  iXLen);
+
+define <vscale x 2 x i1> @intrinsic_vmfne_mask_vf_nxv2bf16_bf16(<vscale x 2 x i1> %0, <vscale x 2 x bfloat> %1, bfloat %2, <vscale x 2 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vf_nxv2bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, mf2, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfne.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x i1> @llvm.riscv.vmfne.mask.nxv2bf16.bf16(
+    <vscale x 2 x i1> %0,
+    <vscale x 2 x bfloat> %1,
+    bfloat %2,
+    <vscale x 2 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 2 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.bf16(
+  <vscale x 4 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfne_vf_nxv4bf16_bf16(<vscale x 4 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, ma
+; CHECK-NEXT:    vmfne.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfne.nxv4bf16.bf16(
+    <vscale x 4 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16(
+  <vscale x 4 x i1>,
+  <vscale x 4 x bfloat>,
+  bfloat,
+  <vscale x 4 x i1>,
+  iXLen);
+
+define <vscale x 4 x i1> @intrinsic_vmfne_mask_vf_nxv4bf16_bf16(<vscale x 4 x i1> %0, <vscale x 4 x bfloat> %1, bfloat %2, <vscale x 4 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vf_nxv4bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m1, ta, mu
+; CHECK-NEXT:    vmv1r.v v10, v0
+; CHECK-NEXT:    vmv1r.v v0, v9
+; CHECK-NEXT:    vmfne.vf v10, v8, fa0, v0.t
+; CHECK-NEXT:    vmv.v.v v0, v10
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x i1> @llvm.riscv.vmfne.mask.nxv4bf16.bf16(
+    <vscale x 4 x i1> %0,
+    <vscale x 4 x bfloat> %1,
+    bfloat %2,
+    <vscale x 4 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 4 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.bf16(
+  <vscale x 8 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfne_vf_nxv8bf16_bf16(<vscale x 8 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, ma
+; CHECK-NEXT:    vmfne.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfne.nxv8bf16.bf16(
+    <vscale x 8 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16(
+  <vscale x 8 x i1>,
+  <vscale x 8 x bfloat>,
+  bfloat,
+  <vscale x 8 x i1>,
+  iXLen);
+
+define <vscale x 8 x i1> @intrinsic_vmfne_mask_vf_nxv8bf16_bf16(<vscale x 8 x i1> %0, <vscale x 8 x bfloat> %1, bfloat %2, <vscale x 8 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vf_nxv8bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m2, ta, mu
+; CHECK-NEXT:    vmv1r.v v11, v0
+; CHECK-NEXT:    vmv1r.v v0, v10
+; CHECK-NEXT:    vmfne.vf v11, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v11
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x i1> @llvm.riscv.vmfne.mask.nxv8bf16.bf16(
+    <vscale x 8 x i1> %0,
+    <vscale x 8 x bfloat> %1,
+    bfloat %2,
+    <vscale x 8 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 8 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.bf16(
+  <vscale x 16 x bfloat>,
+  bfloat,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfne_vf_nxv16bf16_bf16(<vscale x 16 x bfloat> %0, bfloat %1, iXLen %2) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, ma
+; CHECK-NEXT:    vmfne.vf v0, v8, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfne.nxv16bf16.bf16(
+    <vscale x 16 x bfloat> %0,
+    bfloat %1,
+    iXLen %2)
+
+  ret <vscale x 16 x i1> %a
+}
+
+declare <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16(
+  <vscale x 16 x i1>,
+  <vscale x 16 x bfloat>,
+  bfloat,
+  <vscale x 16 x i1>,
+  iXLen);
+
+define <vscale x 16 x i1> @intrinsic_vmfne_mask_vf_nxv16bf16_bf16(<vscale x 16 x i1> %0, <vscale x 16 x bfloat> %1, bfloat %2, <vscale x 16 x i1> %3, iXLen %4) nounwind {
+; CHECK-LABEL: intrinsic_vmfne_mask_vf_nxv16bf16_bf16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, a0, e16alt, m4, ta, mu
+; CHECK-NEXT:    vmv1r.v v13, v0
+; CHECK-NEXT:    vmv1r.v v0, v12
+; CHECK-NEXT:    vmfne.vf v13, v8, fa0, v0.t
+; CHECK-NEXT:    vmv1r.v v0, v13
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x i1> @llvm.riscv.vmfne.mask.nxv16bf16.bf16(
+    <vscale x 16 x i1> %0,
+    <vscale x 16 x bfloat> %1,
+    bfloat %2,
+    <vscale x 16 x i1> %3,
+    iXLen %4)
+
+  ret <vscale x 16 x i1> %a
+}
+