[Mlir-commits] [mlir] [MLIR][Linalg] Introduce Python API for linalg.batch_matmul Ops. (PR #127614)

[Renato Golin]

================
@@ -1497,7 +1497,7 @@ func.func @matmul_transpose_b(%arg0: memref<3x5xf32>, %arg1: memref<7x5xf32>, %a
 // CHECK-SAME:                                    %[[VAL_0:.*]]: memref<5xf32>,
 // CHECK-SAME:                                    %[[VAL_1:.*]]: memref<2x5x7xf32>,
 // CHECK-SAME:                                    %[[VAL_2:.*]]: memref<2x3x7xf32>) {
-// CHECK:           linalg.batch_matmul ins(%[[VAL_0]], %[[VAL_1]] : memref<5xf32>, memref<2x5x7xf32>) outs(%[[VAL_2]] : memref<2x3x7xf32>) indexing_maps = [#[[$ATTR_0]], #[[$ATTR_1]], #[[$ATTR_2]]]
+// CHECK:           linalg.batch_matmul indexing_maps = [#[[$ATTR_0]], #[[$ATTR_1]], #[[$ATTR_2]]] ins(%[[VAL_0]], %[[VAL_1]] : memref<5xf32>, memref<2x5x7xf32>) outs(%[[VAL_2]] : memref<2x3x7xf32>)
----------------
rengolin wrote:

Why did the syntax change? This is already in ODS, so the Python interface should make no difference here.

https://github.com/llvm/llvm-project/pull/127614