[PATCH] D87972: [OldPM] Pass manager: run SROA after (simple) loop unrolling

[Michael Liao via Phabricator via cfe-commits]

hliao added a comment.

In D87972#2284096 <https://reviews.llvm.org/D87972#2284096>, @lebedev.ri wrote:

> In D87972#2284064 <https://reviews.llvm.org/D87972#2284064>, @xbolva00 wrote:
>
>> In D87972#2284060 <https://reviews.llvm.org/D87972#2284060>, @MaskRay wrote:
>>
>>> I have tested this patch internally and seen gains and losses. On one document search related benchmark 3~5% improvement. One zippy (snappy) there is 3~5% regression. Perhaps we do need a conditional extra SROA run.
>>
>> Snappy  - you mean public https://github.com/google/snappy?
>>
>> Well, it should be possible to analyze it...
>
>
>
>> @lebedev.ri any perf data from testsuite/rawspeed?
>
> I did look. F13016699: sroa-after-unroll.rsbench.txt <https://reviews.llvm.org/F13016699>
> This suggests that geomean is `-0.8%` runtime improvement,
> with ups&downs. F13016722: image.png <https://reviews.llvm.org/F13016722>
>
> But as i have said in the patch's description, i stumbled into this when writing new code, where the effect is //much// larger.

We probably need to collect more performance data of more benchmarks on more platforms (different targets) to understand the impact. I hesitate to add https://reviews.llvm.org/rG1f4e7463b5e3ff654c84371527767830e51db10d as a generic one as some targets may have regressions due to potentially very different memory access patterns.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D87972/new/

https://reviews.llvm.org/D87972