[PATCH] D158613: [AArch64] Mark known zero for high 16-bits of uaddlv intrinsic output with v8i8

JinGu Kang via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue Aug 29 02:22:35 PDT 2023


jaykang10 added a comment.

Thanks for comments.

> Can we also handle v16i8?

Yep, let me add the type.

> Is it worth trying to use a more precise bound, instead of just the approximate power of two?  The largest number of bits that can actually be set for a v8i8 is 11 (the number of bits set in 8*255).  If we're using a more precise bound like this, we could also handle other forms of uaddlv.

Yep, I think it would be good to make more bits zero. Let me update the bound.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D158613/new/

https://reviews.llvm.org/D158613



More information about the llvm-commits mailing list