[all-commits] [llvm/llvm-project] 10c54d: [AMDGPU][HWEvents] Refactor VMEM_ACCESS as VMEM_RE...

[Pierre van Houtryve via All-commits]

  Branch: refs/heads/users/pierre-vh/simplify-vmemaccess-event
  Home:   https://github.com/llvm/llvm-project
  Commit: 10c54d2751d51a8f120893dd51b1deac23a898ef
      https://github.com/llvm/llvm-project/commit/10c54d2751d51a8f120893dd51b1deac23a898ef
  Author: pvanhout <pierre.vanhoutryve at amd.com>
  Date:   2026-06-23 (Tue, 23 Jun 2026)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUHWEvents.cpp
    M llvm/lib/Target/AMDGPU/AMDGPUHWEvents.def
    M llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp

  Log Message:
  -----------
  [AMDGPU][HWEvents] Refactor VMEM_ACCESS as VMEM_READ_ACCESS

Instead of having an HWEvent that can be either a read or a write
depending on the target, keep the events as straightforward as
possible and let InsertWaitCnt interpret it. Rename VMEM_ACCESS
to VMEM_READ_ACCESS and set VMEM_STORE_ACCESS & similar events
even if the target does not have a VSCnt.

I think this conceptually makes more sense.
This separates concerns better so that HWEvents nodels events
objectively, and InsertWaitCnt handles them as necessary for the task
it is trying to achieve (insert wait instructions).


  Commit: df64d085e9b7bffbbf7383ff207a09b2e69f9351
      https://github.com/llvm/llvm-project/commit/df64d085e9b7bffbbf7383ff207a09b2e69f9351
  Author: pvanhout <pierre.vanhoutryve at amd.com>
  Date:   2026-06-23 (Tue, 23 Jun 2026)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp

  Log Message:
  -----------
  Adjust comment


Compare: https://github.com/llvm/llvm-project/compare/f6589a17b507...df64d085e9b7

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