[PATCH] D152399: [CodeGen] Fine tune MachineFunctionSplitPass (MFS) for FSAFDO.

Wed Jul 5 15:28:12 PDT 2023

xur added a comment.

In D152399#4404698 <https://reviews.llvm.org/D152399#4404698>, @wenlei wrote:

>> we see a 1.2% performance improvement in clang benchmark, 0.9% QPS improvement in our internal search benchmark, and 3%-5% improvement in internal storage benchmark.
>
> This looks promising. We saw improvements from MFS when used with CSSPGO too, but the improvements were smaller than the numbers here. You are not using Propeller or BOLT in your perf testing above, right?
>
> What is the baseline in your experiments? are you comparing 1) FS-AFDO w/ MFS vs FS-AFDO w/o MFS, all using this change; or comparing 2) after vs before this patch, using FS-AFDO + MFS on both sides?
>
> + at hoy @wlei
>
>> ! In D152399#4407001 <https://reviews.llvm.org/D152399#4407001>, @shenhan wrote:
>>  What is the baseline in your experiments? are you comparing 1) FS-AFDO w/ MFS vs FS-AFDO w/o MFS, all using this change; or comparing 2) after vs before this patch, using FS-AFDO + MFS on both sides?
>
> Thanks. I was comparing "1) FS-AFDO w/ MFS vs FS-AFDO w/o MFS". And FS-AFDO profile is collected via -improved-fs-discriminator=true.

We enabled unique-internal-linkage-names. Also we used the iterative AFDO (our default release model): which is the FS-AFDO profile was generated from a previous AFDO built binary. 
The performance gain can varies a bit with different profiles. For clang, it also varies a bit from source version to source versions -- we have seen 0.8% to 1.2%.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152399/new/

https://reviews.llvm.org/D152399