[PATCH] D133770: AMDGPU: Use GlobalPriority for largest register tuples

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: AMDGPU, rampitec, foad.
Herald added subscribers: kosarev, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

Only do this for 16 and 32 register tuples, although we might want to
extend to 8 tuples.

      

It's incredibly expensive to spill these, and doing so majorly
interferes with the ability to allocate anything else in the function.

      

The lit tests show mostly sizeable improvements with a handful of tiny
regressions with large vectors.


https://reviews.llvm.org/D133770

Files:
  llvm/lib/Target/AMDGPU/SIRegisterInfo.td
  llvm/test/CodeGen/AMDGPU/GlobalISel/extractelement-stack-lower.ll
  llvm/test/CodeGen/AMDGPU/GlobalISel/extractelement.i128.ll
  llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.large.ll
  llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll
  llvm/test/CodeGen/AMDGPU/insert_vector_dynelt.ll
  llvm/test/CodeGen/AMDGPU/llvm.amdgcn.iglp.opt.ll
  llvm/test/CodeGen/AMDGPU/load-constant-i16.ll
  llvm/test/CodeGen/AMDGPU/mfma-no-register-aliasing.ll
  llvm/test/CodeGen/AMDGPU/tuple-allocation-failure.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D133770.459719.patch
Type: text/x-patch
Size: 274671 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220913/2f4bf8a4/attachment-0001.bin>