[llvm] [AMDGPU][True16][CodeGen] use vgpr16 for zext patterns (#153894) (PR #154211)

[Brox Chen via llvm-commits]

broxigarchen wrote:

Hi @jplehr and @Kewen12, I've manually tried this patch with the hip docker and it succssfully pass the previous bad point. However I am not able to fully run the validation since I hit some other clang env issue when linking. Is there a way to run the CI with this patch in the hip test? Thanks!

https://github.com/llvm/llvm-project/pull/154211