[PATCH] D146494: [X86] Combine constant vector inputs for FMA

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:54037
+    if (Subtarget.hasAVX2() && V.hasOneUse() &&
+        ISD::isConstantSplatVector(V.getNode(), SplatValue)) {
+      SmallVector<SDValue, 8> Ops;
----------------
RKSimon wrote:
> pengfei wrote:
> > Why just limit to splat constant?
> Also - why the AVX2 and hasOneUse limits?
Sorry - the hasOneUse is probably necessary to stop ping-ponging between negated constants


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D146494/new/

https://reviews.llvm.org/D146494