[Mlir-commits] [mlir] [mlir][vector] Refactor `createWriteOrMaskedWrite` (PR #138137)

[Diego Caballero]

================
@@ -800,7 +802,9 @@ func.func @test_vectorize_dynamic_pack(%arg0: tensor<?x?xf32>, %arg1: tensor<?x?
 //  CHECK-DAG: %[[c16:.*]] = arith.constant 16 : index
 //  CHECK-DAG: %[[c2:.*]] = arith.constant 2 : index
 //  CHECK-DAG: %[[empty:.*]] = tensor.empty(%[[d0]], %[[d1]]) : tensor<?x?x16x2xf32>
-//      CHECK: %[[mask_0:.*]] = vector.create_mask %[[d0]], %[[d1]], %[[c16]], %[[c2]] : vector<4x1x16x2xi1>
+//  CHECK-DAG: %[[d2:.*]] = tensor.dim %[[empty]], {{.*}} : tensor<?x?x16x2xf32>
+//  CHECK-DAG: %[[d3:.*]] = tensor.dim %[[empty]], {{.*}} : tensor<?x?x16x2xf32>
+//      CHECK: %[[mask_0:.*]] = vector.create_mask %[[d2]], %[[d3]], %[[c16]], %[[c2]] : vector<4x1x16x2xi1>
 //      CHECK: %[[masked_write:.*]] = vector.mask %[[mask_0]] {
 // CHECK-SAME:   vector.transfer_write %[[transpose]], %[[empty]][%[[c0_2]], %[[c0_2]], %[[c0_2]], %[[c0_2]]]
----------------
dcaballe wrote:

Tests are correct so it looks like the issue is just in the doc.

https://github.com/llvm/llvm-project/pull/138137