[PATCH] D66071: [X86] Teach lowerV4I32Shuffle to only use broadcasts if the mask has more than one undef element. Prioritize shifts over broadcast in lowerV8I16Shuffle.

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

Are you asking about v8i16 where I switched the shift priority?

I wish we knew the original VT. In the motivating case the shuffle has been widened. So there’s a bitcast that makes the possibility of folding unlikely.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D66071/new/

https://reviews.llvm.org/D66071