[PATCH] D118644: [x86] invert a vector select IR canonicalization with a binop identity constant

[Phoebe Wang via Phabricator via llvm-commits]

pengfei added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:48984-48987
+      case ISD::FADD: // X + -0.0 --> X
+        return C->isZero() && C->isNegative();
+      case ISD::FSUB: // X - 0.0 --> X
+        return C->isZero() && !C->isNegative();
----------------
Do we need to consider the nsz case?


================
Comment at: llvm/test/CodeGen/X86/vector-bo-select.ll:375-376
 ; AVX2-NEXT:    vinserti128 $1, %xmm2, %ymm3, %ymm2
-; AVX2-NEXT:    vpand %ymm1, %ymm2, %ymm1
-; AVX2-NEXT:    vsubps %ymm1, %ymm0, %ymm0
+; AVX2-NEXT:    vsubps %ymm1, %ymm0, %ymm1
+; AVX2-NEXT:    vblendvps %ymm2, %ymm1, %ymm0, %ymm0
 ; AVX2-NEXT:    retq
----------------
Is the left code better?


================
Comment at: llvm/test/CodeGen/X86/vector-bo-select.ll:383-384
 ; AVX512F-NEXT:    kmovw %edi, %k1
-; AVX512F-NEXT:    vmovaps %zmm1, %zmm1 {%k1} {z}
-; AVX512F-NEXT:    vsubps %ymm1, %ymm0, %ymm0
+; AVX512F-NEXT:    vsubps %ymm1, %ymm0, %ymm1
+; AVX512F-NEXT:    vblendmps %zmm1, %zmm0, %zmm0 {%k1}
+; AVX512F-NEXT:    # kill: def $ymm0 killed $ymm0 killed $zmm0
----------------
The left seems better.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D118644/new/

https://reviews.llvm.org/D118644