[all-commits] [llvm/llvm-project] d9012b: [AMDGPU] Cache results in a bitvector for VOPDPair...

[Igor Wodiany via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: d9012b6f605589ceb81b75298cda0bd38d309d81
      https://github.com/llvm/llvm-project/commit/d9012b6f605589ceb81b75298cda0bd38d309d81
  Author: Igor Wodiany <igor.wodiany at amd.com>
  Date:   2026-06-08 (Mon, 08 Jun 2026)

  Changed paths:
    M llvm/lib/Target/AMDGPU/GCNVOPDUtils.cpp

  Log Message:
  -----------
  [AMDGPU] Cache results in a bitvector for VOPDPairingMutation (#201837)

With this change the code no longer checks `(I, J)` pair when it is
known that `(_, J)` is not a valid VOPD instruction. The saving is
achieved by precomputing `(_, Y)` into a bitvector, that is also used by
`(_, I)` check, so the `(_, J)` check is "free".

---------

Co-authored-by: Antonio Carpio <antonio.carpio at amd.com>



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