[PATCH] D153925: [AMDGPU] NFC: Add schedule-relaxed-occupancy to relax occupancy targets for wave-limited/membound kernels

[Jeffrey Byrnes via Phabricator via llvm-commits]

jrbyrnes created this revision.
jrbyrnes added reviewers: kerbowa, rampitec, nhaehnle.
Herald added subscribers: foad, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl, arsenm.
Herald added a project: All.
jrbyrnes requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.

Default scheduling behavior for these types of kernels is to chase high occupancy goals with scheduling heuristics, but allow occupancy drops if we are unable to reach the target.

This (experimental, off-by-default) feature relaxes occupancy target from the beginning, which enables scheduler to produce better ILP schedules.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D153925

Files:
  llvm/lib/Target/AMDGPU/GCNSchedStrategy.cpp
  llvm/test/CodeGen/AMDGPU/schedule-relaxed-occupancy.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D153925.535152.patch
Type: text/x-patch
Size: 42342 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230627/46498f9f/attachment-0001.bin>