[all-commits] [llvm/llvm-project] cb053c: [AMDGPU] Fix gfx1250 WMMA latencies

[Austin Kerbow via All-commits]

  Branch: refs/heads/users/kerbowa/amdgpu-gfx1250-wmma-latencies
  Home:   https://github.com/llvm/llvm-project
  Commit: cb053c1486d9c6c838ac06183bd1a803bae75488
      https://github.com/llvm/llvm-project/commit/cb053c1486d9c6c838ac06183bd1a803bae75488
  Author: Austin Kerbow <Austin.Kerbow at amd.com>
  Date:   2026-06-08 (Mon, 08 Jun 2026)

  Changed paths:
    M llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
    M llvm/lib/Target/AMDGPU/SISchedule.td
    M llvm/test/CodeGen/AMDGPU/misched-into-wmma-hazard-shadow.mir
    M llvm/test/CodeGen/AMDGPU/wmma-coexecution-valu-hazards.mir
    M llvm/test/CodeGen/AMDGPU/wmma-hazards-gfx1250-w32.mir
    A llvm/test/tools/llvm-mca/AMDGPU/gfx1250-wmma-cycles.s

  Log Message:
  -----------
  [AMDGPU] Fix gfx1250 WMMA latencies

- 16x16x64 FP8/BF8 WMMA run in 4 cycles, not 8. Add a 4-cycle
  WriteXDL1PassWMMA write and split the FP8/BF8 InstRW by shape so that
  16x16x128 FP8/BF8 (and F16/BF16) keep the 8-cycle latency.
- f8f6f4 WMMA run in 4 cycles when both matrix inputs are f4 and 8 cycles
  when any input is f6 or f8, instead of the previous 8/16. Update the
  scheduling predicate accordingly.



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