[PATCH] D88882: [AMDGPU] Prefer SplitVectorLoad/Store over expandUnalignedLoad/Store.

[Mirko Brkusanin via Phabricator via llvm-commits]

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

ExpandUnalignedLoad/Store can sometimes cause unnecessary spilling. We 
should prefer splitting vectors if possible.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D88882

Files:
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/test/CodeGen/AMDGPU/load-local-spill.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D88882.296390.patch
Type: text/x-patch
Size: 6189 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20201006/ee9b4095/attachment.bin>