[all-commits] [llvm/llvm-project] cc21aa: [X86] lower1BitShuffle - fold permute(setcc(x, y)) ...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: cc21aa1922b3d0c4fde52046d8d16d1048f8064e
      https://github.com/llvm/llvm-project/commit/cc21aa1922b3d0c4fde52046d8d16d1048f8064e
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2024-01-10 (Wed, 10 Jan 2024)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp
    M llvm/test/CodeGen/X86/pr77459.ll
    M llvm/test/CodeGen/X86/vector-shuffle-v1.ll

  Log Message:
  -----------
  [X86] lower1BitShuffle - fold permute(setcc(x,y)) -> setcc(permute(x),permute(y)) for 32/64-bit element vectors

Noticed in #77459 - for wider element types, its usually better to pre-shuffle the comparison arguments if we can, like we already for broadcasts