[all-commits] [llvm/llvm-project] 491d3d: [X86] combineINSERT_SUBVECTOR - fold insert_subvec...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 491d3dfc761e3a03c6bd187533f4684d6864a8cb
      https://github.com/llvm/llvm-project/commit/491d3dfc761e3a03c6bd187533f4684d6864a8cb
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2025-03-27 (Thu, 27 Mar 2025)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp
    M llvm/test/CodeGen/X86/any_extend_vector_inreg_of_broadcast.ll
    M llvm/test/CodeGen/X86/any_extend_vector_inreg_of_broadcast_from_memory.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i32-stride-5.ll

  Log Message:
  -----------
  [X86] combineINSERT_SUBVECTOR - fold insert_subvector(base,extract_subvector(broadcast)) -> blend shuffle(base,broadcast) (#133083)

If the broadcast is already the full vector width, try to prefer a blend over a vector insertion which is usually a lower latency (and sometimes a lower uop count).



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications