[all-commits] [llvm/llvm-project] a2a73d: AMDGPU: Fix no return atomicrmw fadd v2f16 selecti...

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a2a73d892a3fd16a1dd2218cf142c7e99cc69ba6
      https://github.com/llvm/llvm-project/commit/a2a73d892a3fd16a1dd2218cf142c7e99cc69ba6
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-06-27 (Thu, 27 Jun 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/FLATInstructions.td
    M llvm/lib/Target/AMDGPU/SIISelLowering.cpp
    M llvm/test/CodeGen/AMDGPU/buffer-fat-pointer-atomicrmw-fadd.ll
    M llvm/test/CodeGen/AMDGPU/global-atomicrmw-fadd.ll
    M llvm/test/Transforms/AtomicExpand/AMDGPU/expand-atomic-rmw-fadd.ll

  Log Message:
  -----------
  AMDGPU: Fix no return atomicrmw fadd v2f16 selection for gfx908 (#96948)

We previously would always expand this with a cmpxchg loop, while
it should be the same conditions as the f32 case (except for the
denormal concern).



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