[llvm] [AMDGPU] Enable WMMA256bInsts + Wave32 for gfx1200/gfx1201 + SISchedule fix + TargetParser gfx1200 propagation (PR #202093)

[Alexandra Munkes via llvm-commits]

https://github.com/Malexandra-de commented:

I did check for existing tests, and all of the tests added by this PR have been in LLVM since early 2024, along with the code generation facilities for WMMA/SWMMA instructions. I linked the relevant commit. The proposed SIScheduler changes will break code generation for gfx1200 and gfx1201, so if you wish to still add them, I'd advise to rework them.

https://github.com/llvm/llvm-project/pull/202093