[PATCH] D32451: Improve profile-guided heuristics to use estimated trip count.

[Wei Mi via llvm-commits]

On Wed, Jun 14, 2017 at 10:16 AM, Ayal Zaks via Phabricator
<reviews at reviews.llvm.org> wrote:
> Ayal added a comment.
>
> This has been conceptually approved, but is pending:
>
>> wmi/dehao, any chance you could check if this causes regressions for us?
>
> Any progress?  The follow-up thought raised above deserves a follow-up patch, also pending.

Sorry, our testing system is not quite stable recently so that I get
the result with some delay.  I test our internal benchmarks with
autofdo and I don't find any real regression. Only one regression is
caused by side-effect. I do see some improvements for three benchmarks
but I havn't got chance to look at whether they are real or just side
effect too.

>From our benchmarks result, the patch looks good.

Thanks,
Wei.

>
>
> https://reviews.llvm.org/D32451
>
>
>