[PATCH] D46073: [DagCombine] De Morgan laws: 'nand' logic with an inverted operand

[Roman Lebedev via Phabricator via llvm-commits]

lebedev.ri added a comment.

In https://reviews.llvm.org/D46073#1085463, @spatel wrote:

> In https://reviews.llvm.org/D46073#1080839, @lebedev.ri wrote:
>
> > - Rebased ontop of revised tests.
> > - Handle `nor` pattern too.
> > - Still not sure how to handle `vandnpd`/`vandnps` breakage.
>
>
> I lost track of where we are in the masked merge odyssey. :)


We are *almost* there..

> I think this is the next step

I think we could handle https://reviews.llvm.org/D46031 first.

> but we have to avoid regressing the x86 andnps tests?

Yes.

> Can we use TLI.hasAndNot() to predicate the transform? Ie, if you have 'andn', then there's no point trying to eliminate a 'not' that will get merged with another instruction?




Repository:
  rL LLVM

https://reviews.llvm.org/D46073