[PATCH] D155384: [amdgpu][wip] Precise existing usage calculation in PromoteAllocaToLDS

[Jon Chesterfield via Phabricator via llvm-commits]

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

WIP because test coverage is scarce. Known bug is dynamic LDS only
allocated in kernels will be missing abs metadata and from the frame
attribute.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D155384

Files:
  llvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp
  llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
  llvm/test/CodeGen/AMDGPU/promote-alloca-padding-size-estimate.ll
  llvm/test/CodeGen/AMDGPU/promote-alloca-to-lds-constantexpr-use.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D155384.540744.patch
Type: text/x-patch
Size: 14614 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230715/235114ee/attachment.bin>