[PATCH] D80386: [AMDGPU] Promote alloca to vector in opt

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec created this revision.
rampitec added a reviewer: arsenm.
Herald added subscribers: kerbowa, zzheng, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

Promote alloca to vector before SROA and loop unroll. If we manage
to eliminate allocas before unroll we may choose to unroll less.


https://reviews.llvm.org/D80386

Files:
  llvm/lib/Target/AMDGPU/AMDGPU.h
  llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
  llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp
  llvm/lib/Target/AMDGPU/sroa-before-unroll.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D80386.265534.patch
Type: text/x-patch
Size: 6330 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200521/50381b25/attachment.bin>