[PATCH] D15232: [InstCombine] Aggressively fold compares that can be discovered to be constant

[James Molloy via llvm-commits]

Hi Hal,

Firstly, sorry about the conflicting message asking Philip to review the
patch - I sent it from Phab before I'd seen you'd replied 15 minutes before!

I went to write a better testcase, and actually found that you're right,
-O3 does indeed manage to constant propagate correctly. This is strange,
and I'm confused as to why I didn't see this in any of my testing. I
suspect that I may have over tuned my testcase, or only been testing
specific passes.

Either way, SCCP is the pass that does the magic. It works fine on both my
reduced and my original testcase, so the only thing for me to do is to add
SCCP to the list of passes run late in the LTO pass manager, which I'll do
in another review.

Thankyou for the extra prod. It has avoided a bunch of useless code being
committed!

Cheers,

James

On Thu, 10 Dec 2015 at 15:37 James Molloy via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

> jmolloy added a comment.
>
> Hi Philip,
>
> Would you feel comfortable reviewing the rest of the code? I'm aware that
> David is rather swamped with reviews and am trying to find alternative
> reviewers.
>
> Cheers,
>
> James
>
>
> http://reviews.llvm.org/D15232
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20151210/ab9f053f/attachment.html>