[all-commits] [llvm/llvm-project] 238a97: [AMDGPU] DS loop wait relaxation -- more test case...

[hidekisaito via All-commits]

  Branch: refs/heads/users/hidekisaito/ds-loop-wait-opt-4
  Home:   https://github.com/llvm/llvm-project
  Commit: 238a970d621ed4b0758d8042ec30ed89895c4c3c
      https://github.com/llvm/llvm-project/commit/238a970d621ed4b0758d8042ec30ed89895c4c3c
  Author: Hideki Saito <hidekido at amd.com>
  Date:   2025-12-11 (Thu, 11 Dec 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp
    M llvm/test/CodeGen/AMDGPU/waitcnt-loop-ds-opt-eligible.mir
    A llvm/test/CodeGen/AMDGPU/waitcnt-loop-ds-opt-no-improvement.mir
    A llvm/test/CodeGen/AMDGPU/waitcnt-loop-ds-opt-same-iter-overwrite.mir
    A llvm/test/CodeGen/AMDGPU/waitcnt-loop-ds-opt-same-iter-use.mir
    A llvm/test/CodeGen/AMDGPU/waitcnt-loop-ds-opt-tensor-load.mir

  Log Message:
  -----------
  [AMDGPU] DS loop wait relaxation -- more test cases and improvements to handle them (4/4)

Add handling for same-iteration use/overwrite of DS load results:
- Track DS load destinations and detect when results are used or
  overwritten within the same iteration
- Compute FloorWaitCount for WMMAs that only use flushed loads
Add bailout for tensor_load_to_lds and LDS DMA writes after barrier
Add negative test based on profitability criteria

Assisted-by: Cursor / claude-4.5-opus-high



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