[all-commits] [llvm/llvm-project] af6fa7: [AMDGPU] Add DAG mutation to improve scheduling be...

[Carl Ritson via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: af6fa77a351e64563ef8abe4af2ab65e6aa669fa
      https://github.com/llvm/llvm-project/commit/af6fa77a351e64563ef8abe4af2ab65e6aa669fa
  Author: Carl Ritson <carl.ritson at amd.com>
  Date:   2025-10-21 (Tue, 21 Oct 2025)

  Changed paths:
    A llvm/lib/Target/AMDGPU/AMDGPUBarrierLatency.cpp
    A llvm/lib/Target/AMDGPU/AMDGPUBarrierLatency.h
    M llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
    M llvm/lib/Target/AMDGPU/CMakeLists.txt
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.update.dpp.ll
    A llvm/test/CodeGen/AMDGPU/schedule-barrier-latency.mir
    M llvm/test/CodeGen/AMDGPU/waitcnt-vscnt.ll

  Log Message:
  -----------
  [AMDGPU] Add DAG mutation to improve scheduling before barriers (#142716)

Add scheduler DAG mutation to add data dependencies between atomic
fences and preceding memory reads. This allows some modelling of the
impact an atomic fence can have on outstanding memory accesses.

This is beneficial when a fence would cause wait count insertion, as
more instructions will be scheduled before the fence hiding memory
latency.



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