[PATCH] D107170: [InstCombine] canonicalize cmp-of-bitcast-of-vector-cmp to use zero constant

Roman Lebedev via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sat Jul 31 10:53:33 PDT 2021


lebedev.ri added a comment.

In D107170#2918476 <https://reviews.llvm.org/D107170#2918476>, @spatel wrote:

> In D107170#2918427 <https://reviews.llvm.org/D107170#2918427>, @lebedev.ri wrote:
>
>> Hm, to be noted, this is still only a subpattern, with further generalization being: (https://alive2.llvm.org/ seems to be down)
>
> Thanks! I'll put a TODO note on this. At first glance, I'm not sure what steps would be needed in the more generalized forms. We do have foldBitCastBitwiseLogic(), but there's a comment in there about changing ops from scalar <-> vector.

Mainly the next step is: there shouldn't be `isFreeToInvert()`, it should be `InstCombineInverter` (a.k.a `InstCombineNegator`, but for inversion)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D107170/new/

https://reviews.llvm.org/D107170



More information about the llvm-commits mailing list