[all-commits] [llvm/llvm-project] 579879: [InstCombine] Allow simplifying FP selects of cmpx...

[Ricardo Jesus via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 579879104cbc2599ba08913d20676f7c2f0dd781
      https://github.com/llvm/llvm-project/commit/579879104cbc2599ba08913d20676f7c2f0dd781
  Author: Ricardo Jesus <rjj at nvidia.com>
  Date:   2026-02-19 (Thu, 19 Feb 2026)

  Changed paths:
    M llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
    M llvm/test/Transforms/InstCombine/select-cmpxchg.ll

  Log Message:
  -----------
  [InstCombine] Allow simplifying FP selects of cmpxchg instructions. (#181977)

We already simplify selects that test the flag returned by a cmpxchg and
select between the value the cmpxchg loaded and the compare operand.

This patch extends the fold to FP (and vector) compare-exchange
operations, where the compare operand and loaded value are bitcast.



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