[llvm] [LLVM][AMDGPU] AMDGPUInstCombineIntrinsic for *lane intrinsics (PR #99878)

[Nikita Popov via llvm-commits]

nikic wrote:

If UA requires any non-trivial amount of analysis updates/invalidation (basically, anything that goes significantly beyond "remove from the analysis if the instruction is removed") then we cannot use it inside InstCombine.

> For AMDGCN, having UniformityAnalysis inside InstCombine is a _huge_ deal and totally worth pursuing.

Does it have to be inside InstCombine proper, rather than a separate pass that does uniformity based optimizations, and knows how to keep the analysis up to date through transforms?

https://github.com/llvm/llvm-project/pull/99878