[PATCH] D88471: [Passes] Run peeling as part of simple/full loop unrolling.
Roman Lebedev via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Fri Oct 2 04:37:17 PDT 2020
lebedev.ri added a comment.
In D88471#2308297 <https://reviews.llvm.org/D88471#2308297>, @fhahn wrote:
> In D88471#2308260 <https://reviews.llvm.org/D88471#2308260>, @lebedev.ri wrote:
>
>> The patch doesn't seem to apply for me:
>
> Rebased, should apply now.
It does, thank you.
> In terms of runtime impact, I measured the following impact on SPEC2006 on ARM64 with -O3 -flto, outside the noise:
>
> test-suite...006/450.soplex/450.soplex.test -2.4%
> test-suite...6/464.h264ref/464.h264ref.test 1.3%
> test-suite...T2006/473.astar/473.astar.test -0.9%
>
> Presumably the h264ref regression is related to wrong decisions in the cost modeling and probably warrants some further analysis.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D88471/new/
https://reviews.llvm.org/D88471
More information about the llvm-commits
mailing list