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

[Alexandra Munkes via llvm-commits]

================

----------------
Malexandra-de wrote:

(S)WMMAC blocks VALU during execution.

According to AMDs documentation, for gfx1200, (S)WMMA blocks the VALU slots, so there's no potential to issue vector instructions concurrent to WMMA. See [flag explanation](https://github.com/ROCm/amd_matrix_instruction_calculator/blob/2ef91896bcdc4d26624f952e5c905c787cd9bc9e/matrix_calculator.py#L216) and [the entry for one of the SWMMA instructions](https://github.com/ROCm/amd_matrix_instruction_calculator/blob/2ef91896bcdc4d26624f952e5c905c787cd9bc9e/matrix_calculator.py#L2697).

This feature would be new for gfx1250/gfx1251 to the best of my knowledge. 

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