[X86] Replace avx512f integer mul reduction builtins with generic bui…

…ltin D117829 added the generic "__builtin_reduce_mul" which we can use to replace the x86 specific integer mul reduction builtins - internally these were mapping to the same intrinsic already so there are no test changes required. Differential Revision: https://reviews.llvm.org/D125222
llvm · May 9, 2022 · ec6024d · ec6024d
1 parent 33f02de
commit ec6024d
Show file tree

Hide file tree

Showing 3 changed files with 4 additions and 12 deletions.
diff --git a/clang/include/clang/Basic/BuiltinsX86.def b/clang/include/clang/Basic/BuiltinsX86.def
@@ -2010,8 +2010,6 @@ TARGET_BUILTIN(__builtin_ia32_reduce_fmul_ps512, "ffV16f", "ncV:512:", "avx512f"
 TARGET_BUILTIN(__builtin_ia32_reduce_fmul_ph512, "xxV32x", "ncV:512:", "avx512fp16")
 TARGET_BUILTIN(__builtin_ia32_reduce_fmul_ph256, "xxV16x", "ncV:256:", "avx512fp16,avx512vl")
 TARGET_BUILTIN(__builtin_ia32_reduce_fmul_ph128, "xxV8x", "ncV:128:", "avx512fp16,avx512vl")
-TARGET_BUILTIN(__builtin_ia32_reduce_mul_d512, "iV16i", "ncV:512:", "avx512f")
-TARGET_BUILTIN(__builtin_ia32_reduce_mul_q512, "OiV8Oi", "ncV:512:", "avx512f")
 
 // MONITORX/MWAITX
 TARGET_BUILTIN(__builtin_ia32_monitorx, "vvC*UiUi", "n", "mwaitx")

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -14409,12 +14409,6 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
     Builder.getFastMathFlags().setNoNaNs();
     return Builder.CreateCall(F, {Ops[0]});
   }
-  case X86::BI__builtin_ia32_reduce_mul_d512:
-  case X86::BI__builtin_ia32_reduce_mul_q512: {
-    Function *F =
-        CGM.getIntrinsic(Intrinsic::vector_reduce_mul, Ops[0]->getType());
-    return Builder.CreateCall(F, {Ops[0]});
-  }
 
   // 3DNow!
   case X86::BI__builtin_ia32_pswapdsf:

diff --git a/clang/lib/Headers/avx512fintrin.h b/clang/lib/Headers/avx512fintrin.h
@@ -9323,7 +9323,7 @@ static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_add_epi64(__m512
 }
 
 static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_mul_epi64(__m512i __W) {
-  return __builtin_ia32_reduce_mul_q512(__W);
+  return __builtin_reduce_mul((__v8di)__W);
 }
 
 static __inline__ long long __DEFAULT_FN_ATTRS512 _mm512_reduce_and_epi64(__m512i __W) {
@@ -9343,7 +9343,7 @@ _mm512_mask_reduce_add_epi64(__mmask8 __M, __m512i __W) {
 static __inline__ long long __DEFAULT_FN_ATTRS512
 _mm512_mask_reduce_mul_epi64(__mmask8 __M, __m512i __W) {
   __W = _mm512_mask_mov_epi64(_mm512_set1_epi64(1), __M, __W);
-  return __builtin_ia32_reduce_mul_q512(__W);
+  return __builtin_reduce_mul((__v8di)__W);
 }
 
 static __inline__ long long __DEFAULT_FN_ATTRS512
@@ -9388,7 +9388,7 @@ _mm512_reduce_add_epi32(__m512i __W) {
 
 static __inline__ int __DEFAULT_FN_ATTRS512
 _mm512_reduce_mul_epi32(__m512i __W) {
-  return __builtin_ia32_reduce_mul_d512((__v16si)__W);
+  return __builtin_reduce_mul((__v16si)__W);
 }
 
 static __inline__ int __DEFAULT_FN_ATTRS512
@@ -9410,7 +9410,7 @@ _mm512_mask_reduce_add_epi32( __mmask16 __M, __m512i __W) {
 static __inline__ int __DEFAULT_FN_ATTRS512
 _mm512_mask_reduce_mul_epi32( __mmask16 __M, __m512i __W) {
   __W = _mm512_mask_mov_epi32(_mm512_set1_epi32(1), __M, __W);
-  return __builtin_ia32_reduce_mul_d512((__v16si)__W);
+  return __builtin_reduce_mul((__v16si)__W);
 }
 
 static __inline__ int __DEFAULT_FN_ATTRS512