[PATCH] D28810: [X86][SSE] Use MOVMSK for all_of/any_of reduction patterns

[Andrea Di Biagio via Phabricator via llvm-commits]

andreadb accepted this revision.
andreadb added a comment.
This revision is now accepted and ready to land.

In https://reviews.llvm.org/D28810#663514, @RKSimon wrote:

> To avoid the partial register issue, I've updated the result select code to always perform it as i32/i64 followed by a truncation if necessary.


Nice! That also fixes the issue with partial register writes for OR reductions.

The patch looks good to me now.
Thanks Simon!


Repository:
  rL LLVM

https://reviews.llvm.org/D28810