[Mlir-commits] [mlir] [mlir][linalg] Generic to category specialization (PR #184624)

[Andrzej Warzyński]

================
@@ -62,6 +71,17 @@ func.func @op_matmul(%A: tensor<?x?xf32>, %B: tensor<?x?xf32>, %Out: tensor<?x?x
 // CHECK-NOT: linalg.generic
 // CHECK: linalg.matmul ins(%[[A]], %[[B]] : tensor<?x?xf32>, tensor<?x?xf32>) outs(%[[Out]] : tensor<?x?xf32>) -> tensor<?x?xf32>
 
+// CATEGORY-DAG: #[[$MAP_A:.+]] = affine_map<(d0, d1, d2) -> (d0, d2)>
+// CATEGORY-DAG: #[[$MAP_B:.+]] = affine_map<(d0, d1, d2) -> (d2, d1)>
+// CATEGORY-DAG: #[[$MAP_C:.+]] = affine_map<(d0, d1, d2) -> (d0, d1)>
+
+// CATEGORY-LABEL: op_matmul
+// CATEGORY-SAME: %[[A:.+]]: tensor<?x?xf32>, %[[B:.+]]: tensor<?x?xf32>,  %[[Out:.+]]: tensor<?x?xf32>) -> tensor<?x?xf32>
----------------
banach-space wrote:

[nit] Use upper case for FileCheck "vars": 

```suggestion
// CATEGORY-SAME: %[[A:.+]]: tensor<?x?xf32>, %[[B:.+]]: tensor<?x?xf32>,  %[[OUT:.+]]: tensor<?x?xf32>) -> tensor<?x?xf32>
```

I know that you are just following the existing format that's inconsistent, hence marking this as nit/nice-to-have.

https://github.com/llvm/llvm-project/pull/184624