[PATCH] D110995: [X86] combineMulToPMADDWD - handle any pow2 vector type and split to legal types

[Pengfei Wang via Phabricator via llvm-commits]

pengfei added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:44343
+  // With AVX512 but without BWI, we would need to split v32i16.
+  unsigned NumElts = VT.getVectorNumElements();
+  if (NumElts == 1 || !isPowerOf2_32(NumElts) ||
----------------
Nit: Should this be mul 2 directly? Then we just need to check `32 <= NumElts`.
The 16 in comparison looks more like a magic number.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110995/new/

https://reviews.llvm.org/D110995