Skip to content

[AMDGPU] Change encoding of gfx1250 ld_scale #167777

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Nov 13, 2025
Merged

[AMDGPU] Change encoding of gfx1250 ld_scale #167777

merged 1 commit into from
Nov 13, 2025
+346 −337

Conversation

@rampitec
Copy link
Collaborator

@rampitec rampitec commented Nov 12, 2025

The unused field src2 is now encoded to vgpr0 accoring to the
latest guidelines.

Fixes: SWDEV-565846

@rampitec Graphite App
Copy link
Collaborator Author

rampitec commented Nov 12, 2025 •
edited
Loading

This stack of pull requests is managed by Graphite. Learn more about stacking.

@rampitec rampitec mentioned this pull request Nov 12, 2025
Merged
@rampitec rampitec marked this pull request as ready for review November 12, 2025 22:36
@llvmbot
Copy link
Member

llvmbot commented Nov 12, 2025

@llvm/pr-subscribers-backend-amdgpu

Author: Stanislav Mekhanoshin (rampitec)

Changes

The unused field src2 is now encoded to vgpr0 accoring to the
latest guidelines.

Fixes: SWDEV-565846

Patch is 142.56 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/167777.diff

3 Files Affected:

  • (modified) llvm/lib/Target/AMDGPU/VOP3PInstructions.td (+12-3)
  • (modified) llvm/test/MC/AMDGPU/gfx1250_asm_wmma_w32.s (+122-122)
  • (modified) llvm/test/MC/Disassembler/AMDGPU/gfx1250_dasm_wmma_w32.txt (+212-212)
diff --git a/llvm/lib/Target/AMDGPU/VOP3PInstructions.td b/llvm/lib/Target/AMDGPU/VOP3PInstructions.td
index 31d8bce4d0c87..786e75f081e44 100644
--- a/llvm/lib/Target/AMDGPU/VOP3PInstructions.td
+++ b/llvm/lib/Target/AMDGPU/VOP3PInstructions.td
@@ -2218,7 +2218,7 @@ class VOP3PX2e <bits<8> op, bits<8> LdScaleOp, VOP3PWMMA_Profile P> : Enc128, VO
   let Inst{23-16} = LdScaleOp;
   let Inst{40-32} = scale_src0;
   let Inst{49-41} = scale_src1;
-  let Inst{58-50} = 0; // scale src2
+  let Inst{58-50} = 0x100; // scale src2 = vgpr0 (dummy)
   let Inst{59}    = matrix_b_scale{0}; // scale_op_sel_hi(0)
   let Inst{60}    = 0;                 // scale_op_sel_hi(1)
   let Inst{63-61} = {0, matrix_a_scale_fmt{1-0}}; // neg (lo)
@@ -2433,6 +2433,15 @@ multiclass VOP3P_Real_with_name_gfx12<bits<8> op,
                           string asmName = !cast<VOP3P_Pseudo>(NAME).Mnemonic> :
   VOP3P_Real_with_name<GFX12Gen, op, backing_ps_name, asmName>;
 
+multiclass VOP3P_Real_LD_SCALE_gfx1250<bits<8> op> {
+  defvar ps = !cast<VOP3P_Pseudo>(NAME);
+  def _gfx1250 :
+    VOP3P_Real_Gen<ps, GFX1250Gen, ps.Mnemonic>,
+    VOP3Pe_gfx11_gfx12<op, ps.Pfl> {
+      let Inst{58-50} = 0x100; // scale src2 = vgpr0 (dummy)
+    }
+}
+
 defm V_PK_MIN_NUM_F16 : VOP3P_Real_with_name_gfx12<0x1b, "V_PK_MIN_F16", "v_pk_min_num_f16">;
 defm V_PK_MAX_NUM_F16 : VOP3P_Real_with_name_gfx12<0x1c, "V_PK_MAX_F16", "v_pk_max_num_f16">;
 
@@ -2462,8 +2471,8 @@ defm V_FMA_MIX_F32_BF16 : VOP3P_Realtriple<GFX1250Gen, 0x3d>;
 defm V_FMA_MIXLO_BF16   : VOP3P_Realtriple<GFX1250Gen, 0x3e>;
 defm V_FMA_MIXHI_BF16   : VOP3P_Realtriple<GFX1250Gen, 0x3f>;
 
-defm V_WMMA_LD_SCALE_PAIRED_B32   : VOP3P_Real_gfx1250<0x35>;
-defm V_WMMA_LD_SCALE16_PAIRED_B64 : VOP3P_Real_gfx1250<0x3a>;
+defm V_WMMA_LD_SCALE_PAIRED_B32   : VOP3P_Real_LD_SCALE_gfx1250<0x35>;
+defm V_WMMA_LD_SCALE16_PAIRED_B64 : VOP3P_Real_LD_SCALE_gfx1250<0x3a>;
 
 let AssemblerPredicate = isGFX1250Plus in
 def : AMDGPUMnemonicAlias<"v_fma_mix_f32_f16",  "v_fma_mix_f32">;
diff --git a/llvm/test/MC/AMDGPU/gfx1250_asm_wmma_w32.s b/llvm/test/MC/AMDGPU/gfx1250_asm_wmma_w32.s
index febad4f48ddfd..fcfff9ac5b63d 100644
--- a/llvm/test/MC/AMDGPU/gfx1250_asm_wmma_w32.s
+++ b/llvm/test/MC/AMDGPU/gfx1250_asm_wmma_w32.s
@@ -990,442 +990,442 @@ v_wmma_f32_16x16x128_f8f6f4 v[0:7], v[8:23], v[24:39], v[40:47] neg_hi:[0,0,1]
 
 v_wmma_ld_scale_paired_b32 v1, v2
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s1, s2
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s1, s2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x04,0x00,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 s1, s2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x04,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 2, -4
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 2, -4        ; encoding: [0x00,0x00,0x35,0xcc,0x82,0x88,0x01,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 2, -4        ; encoding: [0x00,0x00,0x35,0xcc,0x82,0x88,0x01,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_a_scale:MATRIX_SCALE_ROW0 matrix_b_scale:MATRIX_SCALE_ROW0
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x08,0x35,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x08,0x35,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_reuse ; encoding: [0x00,0x20,0x35,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_reuse ; encoding: [0x00,0x20,0x35,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse ; encoding: [0x00,0x28,0x35,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse ; encoding: [0x00,0x28,0x35,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x35,0xcc,0x00,0x00,0x00,0x08]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x35,0xcc,0x00,0x00,0x00,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_reuse ; encoding: [0x00,0x40,0x35,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_reuse ; encoding: [0x00,0x40,0x35,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse ; encoding: [0x00,0x40,0x35,0xcc,0x00,0x00,0x00,0x08]
+// GFX1250: v_wmma_ld_scale_paired_b32 s0, s0 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse ; encoding: [0x00,0x40,0x35,0xcc,0x00,0x00,0x00,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_a_scale:MATRIX_SCALE_ROW0 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E8 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E8
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x08]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x28]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x2c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x48]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x4c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E8
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2       ; encoding: [0x00,0x00,0x35,0xcc,0x01,0x05,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x01,0x35,0xcc,0x01,0x05,0x02,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x01,0x35,0xcc,0x01,0x05,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x02,0x35,0xcc,0x01,0x05,0x02,0x00]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x02,0x35,0xcc,0x01,0x05,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale_paired_b32 v1, v2 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse ; encoding: [0x00,0x6a,0x35,0xcc,0x01,0x05,0x02,0x28]
+// GFX1250: v_wmma_ld_scale_paired_b32 v1, v2 matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse ; encoding: [0x00,0x6a,0x35,0xcc,0x01,0x05,0x02,0x2c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5]
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[2:3], s[4:5]
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[2:3], s[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x08,0x00,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[2:3], s[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x08,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 2, -4
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 2, -4      ; encoding: [0x00,0x00,0x3a,0xcc,0x82,0x88,0x01,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 2, -4      ; encoding: [0x00,0x00,0x3a,0xcc,0x82,0x88,0x01,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_a_scale:MATRIX_SCALE_ROW0 matrix_b_scale:MATRIX_SCALE_ROW0
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x08,0x3a,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x08,0x3a,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_reuse ; encoding: [0x00,0x20,0x3a,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_reuse ; encoding: [0x00,0x20,0x3a,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse ; encoding: [0x00,0x28,0x3a,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_a_reuse ; encoding: [0x00,0x28,0x3a,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x3a,0xcc,0x00,0x00,0x00,0x08]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x3a,0xcc,0x00,0x00,0x00,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_reuse ; encoding: [0x00,0x40,0x3a,0xcc,0x00,0x00,0x00,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_reuse ; encoding: [0x00,0x40,0x3a,0xcc,0x00,0x00,0x00,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse ; encoding: [0x00,0x40,0x3a,0xcc,0x00,0x00,0x00,0x08]
+// GFX1250: v_wmma_ld_scale16_paired_b64 s[0:1], s[0:1] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_b_reuse ; encoding: [0x00,0x40,0x3a,0xcc,0x00,0x00,0x00,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_a_scale:MATRIX_SCALE_ROW0 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E8 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E8
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x08]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x0c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x28]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x2c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x48]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x4c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E8
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] ; encoding: [0x00,0x00,0x3a,0xcc,0x02,0x09,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x01,0x3a,0xcc,0x02,0x09,0x02,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E5M3 ; encoding: [0x00,0x01,0x3a,0xcc,0x02,0x09,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x02,0x3a,0xcc,0x02,0x09,0x02,0x00]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 ; encoding: [0x00,0x02,0x3a,0xcc,0x02,0x09,0x02,0x04]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse
 // GFX12-ERR: :[[@LINE-1]]:1: error: instruction not supported on this GPU
-// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse ; encoding: [0x00,0x6a,0x3a,0xcc,0x02,0x09,0x02,0x28]
+// GFX1250: v_wmma_ld_scale16_paired_b64 v[2:3], v[4:5] matrix_a_scale:MATRIX_SCALE_ROW1 matrix_b_scale:MATRIX_SCALE_ROW1 matrix_a_scale_fmt:MATRIX_SCALE_FMT_E5M3 matrix_b_scale_fmt:MATRIX_SCALE_FMT_E4M3 matrix_a_reuse matrix_b_reuse ; encoding: [0x00,0x6a,0x3a,0xcc,0x02,0x09,0x02,0x2c]
 // WAVESIZE-ERR: :[[@LINE-3]]:1: error: instruction requires wavesize=32
 
 v_wmma_scale_f32_16x16x128_f8f6f4 v[0:7], v[8:23], v[24:35], v[40:47], v1, v2 matrix_a_fmt:MATRIX_FMT_BF8 matrix...
[truncated]

shiltian
shiltian reviewed Nov 12, 2025
View reviewed changes
llvm/lib/Target/AMDGPU/VOP3PInstructions.td
def _gfx1250 :
VOP3P_Real_Gen<ps, GFX1250Gen, ps.Mnemonic>,
VOP3Pe_gfx11_gfx12<op, ps.Pfl> {
let Inst{58-50} = 0x100; // scale src2 = vgpr0 (dummy)
Copy link
Contributor

@shiltian shiltian Nov 12, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need both?

Copy link
Collaborator Author

@rampitec rampitec Nov 12, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, this one is for v_wmma_ld_scale_pared itself. The VOP3PX2e is for v_wmma_scale instructions.

Base automatically changed from users/rampitec/11-12-_amdgpu_regenerate_gfx1250_wmma_mc_test._nfc to main November 12, 2025 23:18
@rampitec
[AMDGPU] Change encoding of gfx1250 ld_scale
6a7fe86 
The unused field src2 is now encoded to vgpr0 accoring to the
latest guidelines.

Fixes: SWDEV-565846
@rampitec rampitec force-pushed the users/rampitec/11-12-_amdgpu_change_encoding_of_gfx1250_ld_scale branch from ff61bd6 to 6a7fe86 Compare November 12, 2025 23:20
@rampitec rampitec merged commit 500e6d8 into main Nov 13, 2025
10 checks passed
@rampitec rampitec deleted the users/rampitec/11-12-_amdgpu_change_encoding_of_gfx1250_ld_scale branch November 13, 2025 00:03
@llvm-ci
Copy link
Collaborator

llvm-ci commented Nov 13, 2025

LLVM Buildbot has detected a new failure on builder lldb-x86_64-debian running on lldb-x86_64-debian while building llvm at step 6 "test".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/162/builds/35100

Here is the relevant piece of the build log for the reference 
Step 6 (test) failure: build (failure)
...
PASS: lldb-api :: functionalities/thread/concurrent_events/TestConcurrentManyWatchpoints.py (251 of 3316)
PASS: lldb-api :: functionalities/thread/exit_during_expression/TestExitDuringExpression.py (252 of 3316)
PASS: lldb-api :: lang/cpp/unique-types4/TestUniqueTypes4.py (253 of 3316)
PASS: lldb-api :: commands/expression/static-initializers/TestStaticInitializers.py (254 of 3316)
PASS: lldb-api :: tools/lldb-dap/stackTraceMissingModule/TestDAP_stackTraceMissingModule.py (255 of 3316)
PASS: lldb-api :: python_api/type/TestTypeList.py (256 of 3316)
PASS: lldb-api :: functionalities/breakpoint/comp_dir_symlink/TestCompDirSymLink.py (257 of 3316)
PASS: lldb-api :: tools/lldb-dap/invalidated-event/TestDAP_invalidatedEvent.py (258 of 3316)
PASS: lldb-api :: commands/expression/options/TestExprOptions.py (259 of 3316)
PASS: lldb-api :: functionalities/tail_call_frames/ambiguous_tail_call_seq2/TestAmbiguousTailCallSeq2.py (260 of 3316)
FAIL: lldb-api :: tools/lldb-dap/output/TestDAP_output.py (261 of 3316)
******************** TEST 'lldb-api :: tools/lldb-dap/output/TestDAP_output.py' FAILED ********************
Script:
--
/usr/bin/python3 /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/dotest.py -u CXXFLAGS -u CFLAGS --env LLVM_LIBS_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/./lib --env LLVM_INCLUDE_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/include --env LLVM_TOOLS_DIR=/home/worker/2.0.1/lldb-x86_64-debian/build/./bin --arch x86_64 --build-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex --lldb-module-cache-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex/module-cache-lldb/lldb-api --clang-module-cache-dir /home/worker/2.0.1/lldb-x86_64-debian/build/lldb-test-build.noindex/module-cache-clang/lldb-api --executable /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/lldb --compiler /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/clang --dsymutil /home/worker/2.0.1/lldb-x86_64-debian/build/./bin/dsymutil --make /usr/bin/gmake --llvm-tools-dir /home/worker/2.0.1/lldb-x86_64-debian/build/./bin --lldb-obj-root /home/worker/2.0.1/lldb-x86_64-debian/build/tools/lldb --lldb-libs-dir /home/worker/2.0.1/lldb-x86_64-debian/build/./lib --cmake-build-type Release -t /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/tools/lldb-dap/output -p TestDAP_output.py
--
Exit Code: 1

Command Output (stdout):
--
lldb version 22.0.0git (https://github.com/llvm/llvm-project.git revision 500e6d8b2884c27671c7cada645987f4e96dd181)
  clang revision 500e6d8b2884c27671c7cada645987f4e96dd181
  llvm revision 500e6d8b2884c27671c7cada645987f4e96dd181
Skipping the following test categories: ['libc++', 'msvcstl', 'dsym', 'pdb', 'gmodules', 'debugserver', 'objc']

--
Command Output (stderr):
--
Change dir to: /home/worker/2.0.1/lldb-x86_64-debian/llvm-project/lldb/test/API/tools/lldb-dap/output
runCmd: settings clear --all

output: 

runCmd: settings set symbols.enable-external-lookup false

output: 

runCmd: settings set target.inherit-tcc true

output: 

runCmd: settings set target.disable-aslr false

output: 

runCmd: settings set target.detach-on-error false

output:

git-crd pushed a commit to git-crd/crd-llvm-project that referenced this pull request Nov 13, 2025
@rampitec @git-crd
[AMDGPU] Change encoding of gfx1250 ld_scale (llvm#167777)
e098cf9 
The unused field src2 is now encoded to vgpr0 accoring to the
latest guidelines.

Fixes: SWDEV-565846
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@shiltian shiltian shiltian approved these changes

@changpeng changpeng Awaiting requested review from changpeng

Assignees

No one assigned

Labels

backend:AMDGPU

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@rampitec @llvmbot @llvm-ci @shiltian