[PATCH] D67318: [SimplifyCFG] FoldTwoEntryPHINode(): consider *total* speculation cost, not per-BB cost

[James Molloy via llvm-commits]

Tangentially, that fptosi/fadd/sitofp pattern is what the Float2Int pass is
meant to fix up. I wonder if it's legality checks could be extended to
match this benchmark (we need to ensure the int is less than 23 bits).

On Tue, 17 Sep 2019, 09:08 Roman Lebedev via Phabricator, <
reviews at reviews.llvm.org> wrote:

> lebedev.ri added a comment.
>
> In D67318#1671803 <https://reviews.llvm.org/D67318#1671803>, @xbolva00
> wrote:
>
> > Just curious, did you try threshold 5 or 6 for your rawspeed benchmark?
>
>
> No, i have not tried any other thresholds here.
> I'm personally not aware of any patterns that would benefit from those
> thresholds, although i did not look.
>
>
> Repository:
>   rL LLVM
>
> CHANGES SINCE LAST ACTION
>   https://reviews.llvm.org/D67318/new/
>
> https://reviews.llvm.org/D67318
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190917/9112f3e6/attachment.html>