[Mlir-commits] [mlir] [mlir] [XeGPU] Add XeGPU workgroup to subgroup pass (PR #139477)

[Frank Schlimbach]

================
@@ -0,0 +1,66 @@
+// RUN: mlir-opt --xegpu-wg-to-sg -split-input-file %s | FileCheck %s
+
+gpu.module @test_round_robin_assignment {
+  // CHECK: test_create_nd_tdesc
+  // CHECK: %[[ARG_0:.*]]: memref<24x32xf32>
+  gpu.func @test_create_nd_tdesc(%src: memref<24x32xf32>) {
+      // CHECK-COUNT-12: %[[TDESC:.*]] = xegpu.create_nd_tdesc %[[ARG_0]][%{{.*}}, %{{.*}}] : memref<24x32xf32> -> !xegpu.tensor_desc<2x2xf32, #xegpu.layout<lane_layout = [2, 2], lane_data = [1, 1]>>
----------------
fschlimb wrote:

The variable match `%[[TDESC:.*]]` does not really make a difference. Could be replaced by a reg-ex or simply removed.
Same below.

https://github.com/llvm/llvm-project/pull/139477