[PATCH] D90901: [X86] Don't fold (fneg (fma (fneg X), Y, (fneg Z))) to (fma X, Y, Z)

[Qiu Chaofan via Phabricator via llvm-commits]

qiucf added inline comments.


================
Comment at: llvm/test/CodeGen/X86/fma-signed-zero.ll:12
+
+define float @fneg_fma32(float %x, float %y, float %z) {
+; NO-NSZ-OPTION-LABEL: fneg_fma32:
----------------
You can pre-commit (or at least stage them first and do git-diff) this test so that what optimizations are prevented is clear.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D90901/new/

https://reviews.llvm.org/D90901