[llvm] [AMDGPU] Improved Lowering of abs(i8/i16) and -abs(i8/i16) (PR #165626)

[Simon Pilgrim via llvm-commits]

https://github.com/RKSimon commented:

I'm really not keen on overriding expansions like this. Have you tried adding a local combineABS to just extend anything smaller than i32?

https://github.com/llvm/llvm-project/pull/165626