[PATCH] D26583: [X86][SSE] Improve SINT_TO_FP of boolean vector results (signum)

[Simon Pilgrim via llvm-commits]

RKSimon added a comment.

In https://reviews.llvm.org/D26583#594186, @zvi wrote:

> This patch LGTM, but could you please explain this part in the summary? "This is not necessary for AVX512 targets where boolean vectors are legal."
>  Apparently, this also helps AVX512 targets when the vector of i1's is not legal.


AVX512 manages to lower ( sint_to_fp vXi1 ) into some form of ( select mask, 1.0f , 0.0f ) in most cases.


Repository:
  rL LLVM

https://reviews.llvm.org/D26583