[PATCH] D64642: [AMDGPU] Tune inlining parameters for AMDGPU target

[Daniil Fukalov via Phabricator via llvm-commits]

dfukalov created this revision.
dfukalov added reviewers: arsenm, rampitec.
Herald added subscribers: haicheng, hiraditya, eraman, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.
dfukalov added a project: AMDGPU.

Since the target has no significant advantage of vectorization,
vector instructions bous threshold bonus should be optional.

amdgpu-inline-arg-alloca-cost parameter default value and the target
InliningThresholdMultiplier value tuned then respectively.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D64642

Files:
  llvm/include/llvm/Analysis/TargetTransformInfo.h
  llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
  llvm/include/llvm/CodeGen/BasicTTIImpl.h
  llvm/lib/Analysis/InlineCost.cpp
  llvm/lib/Analysis/TargetTransformInfo.cpp
  llvm/lib/Target/AMDGPU/AMDGPUInline.cpp
  llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
  llvm/test/CodeGen/AMDGPU/amdgpu-inline.ll
  llvm/test/Transforms/Inline/AMDGPU/inline-amdgpu-vecbonus.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D64642.209505.patch
Type: text/x-patch
Size: 9019 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190712/fa5dfe63/attachment.bin>