[llvm] [NVPTX] Cleanup and refactor atomic lowering (PR #133781)

[Artem Belevich via llvm-commits]

Artem-B wrote:

Speaking of atomics and vectors, we do have an implementation gap for atomics on 128-bit types like i128 or float4. 
#122760

https://github.com/llvm/llvm-project/pull/133781