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

[Alexandra Munkes via llvm-commits]

================
@@ -494,6 +494,20 @@ def : HWWriteRes<WriteSALUDummy, [HWSALU],  2>;
 
 }  // End SchedModel = GFX12SpeedModel
 
+let SchedModel = GFX12SpeedModel in {
----------------
Malexandra-de wrote:

This will overwrite the existing GFX12SpeedModel, according to the documentation on the TableGen-Syntax:
https://bcain-llvm.readthedocs.io/projects/llvm/en/latest/TableGen/LangIntro/#let-expressions-within-a-record
For GFX1250, there is a separate GFX125X block to handle these latencies.

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