[PATCH] D86888: RegAllocFast: Make self loop live-out heuristic more aggressive

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: qcolombet, MatzeB.
Herald added subscribers: kerbowa, hiraditya, nhaehnle, jvesely.
Herald added a project: LLVM.
arsenm requested review of this revision.
Herald added a subscriber: wdng.

This currently has no impact on code, but prevents sizeable code size
regressions after D52010 <https://reviews.llvm.org/D52010>. This prevents spilling and reloading all
values inside blocks that loop back. Add a baseline test which would
regress without this patch.


https://reviews.llvm.org/D86888

Files:
  llvm/lib/CodeGen/RegAllocFast.cpp
  llvm/test/CodeGen/AMDGPU/fastregalloc-self-loop-heuristic.mir

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D86888.289007.patch
Type: text/x-patch
Size: 10317 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200831/97f6217c/attachment.bin>