[PATCH] D81991: [AMDGPU] Control num waves per EU for implicit work-group size

[Pushpinder Singh via Phabricator via llvm-commits]

pdhaliwal created this revision.
Herald added subscribers: llvm-commits, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl, arsenm, qcolombet.
Herald added a project: LLVM.
pdhaliwal added reviewers: rampitec, b-sumner, sameerds, msearles.

If amdgpu-flat-work-group-size is not specified in LLVM IR, the backend
uses default value of 1024. For this, minimum waves per EU should be 4.
However, backend is still setting minimum value to 1 instead of calculated
value. This is not observed normally as frontend always provide
amdgpu-flat-work-group-size attribute.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D81991

Files:
  llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
  llvm/test/CodeGen/AMDGPU/InlineAsmCrash.ll
  llvm/test/CodeGen/AMDGPU/accvgpr-copy.mir
  llvm/test/CodeGen/AMDGPU/reserve-vgpr-for-sgpr-spill.ll
  llvm/test/CodeGen/AMDGPU/spill-vgpr-to-agpr.ll
  llvm/test/CodeGen/AMDGPU/waitcnt-overflow.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D81991.271275.patch
Type: text/x-patch
Size: 6677 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200617/978c1db7/attachment.bin>