[Mlir-commits] [mlir] [MLIR][XeGPU] Distribute load_gather/store_scatter op from Wg To Sg (PR #154420)

[Adam Siemieniuk]

================
@@ -264,6 +264,47 @@ gpu.module @test_distribution {
   gpu.return
   }
 
+  // CHECK-LABEL: @load_gather
+  // CHECK-SAME: %[[ARG0:.*]]: memref<?xf16>
+  gpu.func @load_gather(%src : memref<?xf16>) {
+    // CHECK: %[[CST:.*]] = arith.constant dense<0> : vector<32x4xindex>
+    // CHECK: %[[MASK:.*]] = arith.constant dense<true> : vector<32x4xi1>
+    // CHECK: %[[LOAD:.*]] = xegpu.load %[[ARG0]][%[[CST]]], %[[MASK]] <{chunk_size = 1 : i64, l1_hint = #xegpu.cache_hint<cached>}> : memref<?xf16>, vector<32x4xindex>, vector<32x4xi1> -> vector<32x4xf16>
----------------
adam-smnk wrote:

nit: if possible, split the check into two lines for readability

https://github.com/llvm/llvm-project/pull/154420