[PATCH] D15537: limit the number of instructions per block examined by dead store elimination

Friedman, Eli via llvm-commits llvm-commits at lists.llvm.org
Mon Aug 22 17:50:45 PDT 2016


On 8/22/2016 11:28 AM, Daniel Berlin via llvm-commits wrote:
> +philip because he expressed concerns around limiting
>
> +eli
>
> Eli, i think your patch to DSE to do exact may-throw tracking is what 
> made this worse.
>
> https://llvm.org/bugs/show_bug.cgi?id=29064

I don't think it's the fault of my recent patch. The stack trace from 
that bug points into MemDep, not my new code, and the patch which fixes 
it doesn't touch my new code.

> I'm not sure we should limit DSE.

An arbitrary limit isn't exactly ideal, but there probably isn't any 
better way to fix the issue given the current algorithm used by DSE and 
MemDep.

-Eli

>
> I have an analysis pass just about ready to do the EBB/maythrow stuff, 
> which gives "some" may-throw tracking, just not as exact as is being 
> done here.
> Rather than limit DSE, i think we should revert the more-exact 
> tracking for now unless we can show it produces significant perf 
> improvements.
>
> Thoughts?
>
>
>
> On Mon, Aug 22, 2016 at 10:59 AM, David Li <davidxl at google.com 
> <mailto:davidxl at google.com>> wrote:
>
>     davidxl added a comment.
>
>     The patch fixes the issue in
>     https://llvm.org/bugs/show_bug.cgi?id=29064
>     <https://llvm.org/bugs/show_bug.cgi?id=29064>
>
>
>     https://reviews.llvm.org/D15537 <https://reviews.llvm.org/D15537>
>
>
>
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits


-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160822/70d7ace2/attachment.html>


More information about the llvm-commits mailing list