[PATCH] D46706: [PM/LoopUnswitch] Support partial trivial unswitching.

[Chandler Carruth via llvm-commits]

On Thu, May 10, 2018 at 5:40 PM Sanjoy Das via Phabricator <
reviews at reviews.llvm.org> wrote:

> sanjoy added a comment.
>
> Actually now that I think of it, I wonder if looking back through the
> condition expression tree to find all loop invariant values is necessary --
> if we see these kinds of cases `(loop_varying & loop_invariant0) &
> loop_invariant1)` then perhaps we should teach LICM et. al. to reassociate
> and hoist the `loop_invariant0 & loop_invariant1` bit instead of worrying
> about them in loop unswitch?
>

I think we should teach LICM to do this if it doesn't already, but it seems
like this could also come up due to prior (trivial) unswitches? So it might
still be important to handle them?

>
>
> Repository:
>   rL LLVM
>
> https://reviews.llvm.org/D46706
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180510/d97d69a7/attachment.html>