[PATCH] D87342: Allow targets to augment computeKnownBits with their analysis using TargetTransformInfo

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

In D87342#2314787 <https://reviews.llvm.org/D87342#2314787>, @qcolombet wrote:

>> If the motivating cases all start with a gep, then could we add the extra logic to a wrapper around the regular computeKnownBits()?
>
> Unfortunately the motivating example doesn't start with a gep, or at least, there is no guarantee it would start with a gep.
>
> Anyway, I can already push the NFC changes like the added tests cases and refactoring of KnownBits.

Sure, having those in place will help to focus on the real diffs.

> In the meantime, any other idea on how we could have GEPs being more precise?
> Should I resurrect https://reviews.llvm.org/D86364?

Ok - I think that's the theoretically/philosophically right way to proceed (target-independent). I don't have any ideas yet on how to limit the cost, but maybe we can collectively find some more savings in an update of that patch.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D87342/new/

https://reviews.llvm.org/D87342