[PATCH] D15393: [X86] Order the local stack symbols to improve code size and locality.

[Gerolf Hoflehner via llvm-commits]

This was just about a test in the llvm test suite, I assume you have run it and not seen any issue (sorry for the confusing link).

MultiSource/Benchmarks/Ptrdist/ks/ks <https://smooshbase.apple.com/perf/db_default/v4/nts/graph?highlight_run=114068&plot.746=313.746.3> (compiled with O3 FLTO).

-Gerolf

> On Feb 22, 2016, at 10:00 AM, Zia Ansari <zia.ansari at intel.com> wrote:
> 
> zansari added a comment.
> 
> In http://reviews.llvm.org/D15393#355355, @Gerolf wrote:
> 
>> Hi Zia,
>> 
>> do you have HSW performance numbers for this change? An internal bot is logging a >10% regression for MultiSource/Benchmarks/Ptrdist/ks/ks https://smooshbase.apple.com/perf/db_default/v4/nts/graph?highlight_run=114068&plot.746=313.746.3 (O3 flto) pinned to this change.
>> 
>> Thanks!
>> Gerolf
> 
> 
> Hi Gerolf,
> 
> I ran internal performance testing on Atom and HSW (spec2k/2006, eembc, and a few other benchmarks), and I didn't see any negative swings. I'll be happy to look into your regression if you can give me something to look at. The link doesn't open for me (page not available).
> 
>> If you can reproduce it perhaps there is a tuning opportunity.
> 
> 
> Is this a test to which I can get access easily? Let me know how I can get/run it and I'll take a look.
> 
> Thanks,
> Zia.
> 
> 
> Repository:
>  rL LLVM
> 
> http://reviews.llvm.org/D15393
> 
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160224/1ac265de/attachment.html>