[Mlir-commits] [mlir] [mlir][linalg][nfc] Fix `linalg.matmul_transpose_a` def. (PR #97690)

[Jerry Shih]

================
@@ -1336,7 +1336,7 @@ structured_op: !LinalgStructuredOpConfig
     name: C
     kind: output_tensor
     type_var: U
-    shape_map: affine_map<()[s0, s1, s2] -> (s2, s1)>
+    shape_map: affine_map<()[s0, s1, s2] -> (s1, s2)>
----------------
JerryShih wrote:

I don't think we have the consistent meaning for s0, s1 and s2. They are just the symbol names.
This `LinalgNamedStructuredOps.yaml` file is generated from `mlir/python/mlir/dialects/linalg/opdsl/ops/core_named_ops.py`.

Here is the document:
https://mlir.llvm.org/docs/Dialects/Linalg/OpDSL/

In this pr, I update the `matmul_transpose_a` def in `core_named_ops.py` to follow the same dim order as `batch_matmul_transpose_a`.

https://github.com/llvm/llvm-project/blob/5fc7342a836cea6409719bdbdb5b69e8e4e7c570/mlir/python/mlir/dialects/linalg/opdsl/ops/core_named_ops.py#L530-L532

https://github.com/llvm/llvm-project/pull/97690