[llvm] [WIP][AMDGPU] combine uniform AMDGPU lane Intrinsics (PR #116953)

[Pankaj Dwivedi via llvm-commits]

PankajDwivedi-25 wrote:

> Also, there must be a test that shows how this pass can be used to eliminate a trivial waterfall loop. That is a loop which depends on "wfall", and all threads in the wave finish their work in the first iteration itself.

I see this case is not getting optimized in the checks generated & the loop structure remains intact.

https://github.com/llvm/llvm-project/pull/116953