[Mlir-commits] [mlir] [MLIR][XeGPU] Add support for cross-subgroup reduction from wg to sg (PR #170936)

[Artem Kroviakov]

================
@@ -634,6 +642,137 @@ gpu.module @test_distribution {
     gpu.return
   }
 
+  // CHECK-LABEL: gpu.func @vector_reduce_cross_sg_dim_1
+  // CHECK-SAME: (%[[ARG0:.*]]: memref<?xf32>)
+  gpu.func @vector_reduce_cross_sg_dim_1(%src: memref<?xf32>) {
+    // CHECK-DAG: %[[CST:.*]] = arith.constant dense<1.000000e+00> : vector<1x32xf32>
+    // CHECK-DAG: %[[CST_0:.*]] = arith.constant dense<0> : vector<1x1x32xindex>
+    // CHECK-DAG: %[[CST_1:.*]] = arith.constant dense<true> : vector<1x1x32xi1>
+    // CHECK-DAG: %[[LOAD:.*]] = xegpu.load %{{.*}}[%[[CST_0]]], %[[CST_1]] <{chunk_size = 1 : i64}> : memref<?xf32>, vector<1x1x32xindex>, vector<1x1x32xi1> -> vector<1x1x32xf32>
+    // CHECK-DAG: %[[CST_2:.*]] = arith.constant dense<0.000000e+00> : vector<1x32xf32>
+    // CHECK-DAG: %[[LOCAL_REDUCE:.*]] = vector.multi_reduction <add>, %[[LOAD]], %[[CST_2]] [1] : vector<1x1x32xf32> to vector<1x32xf32>
+    // CHECK-DAG: %[[SHAPE_CAST:.*]] = vector.shape_cast %[[LOCAL_REDUCE]] : vector<1x32xf32> to vector<1x32xf32>
----------------
akroviakov wrote:

Seems redundant

https://github.com/llvm/llvm-project/pull/170936