[llvm] [NVPTX] Use PRMT instruction to lower i16 bswap (PR #168968)

[via llvm-commits]

Chengjunp wrote:

> If you wanted to go above and beyond, I think it would be best to instead implement this (and all other bswap expansion) as a custom legalization to allow for DAGCombiner to optimize them.

I see. Do you mean that we just lower all the bswap into bit operations first and let DAGCombiner decide how to optimize using PRMT?

https://github.com/llvm/llvm-project/pull/168968