[Mlir-commits] [mlir] [MLIR] Implement emulation of static indexing subbyte type vector stores (PR #115922)

[Andrzej Warzyński]

================
@@ -287,3 +287,93 @@ func.func @vector_maskedload_i4_constant_mask_unaligned(%passthru: vector<5xi2>)
 // CHECK: %[[RESULT:.+]] = vector.extract_strided_slice %[[SELECT]]
 // CHECK-SAME: {offsets = [1], sizes = [5], strides = [1]} : vector<8xi2> to vector<5xi2>
 // CHECK: return %[[RESULT]] : vector<5xi2>
+
+///----------------------------------------------------------------------------------------
+/// vector.store
+///----------------------------------------------------------------------------------------
+
+// -----
+
+func.func @vector_store_i2_atomic(%arg0: vector<7xi2>) {
+    %0 = memref.alloc() : memref<3x7xi2>
+    %c0 = arith.constant 0 : index
+    %c1 = arith.constant 1 : index
+    vector.store %arg0, %0[%c1, %c0] :memref<3x7xi2>, vector<7xi2>
+    return
+}
+
+// in this example, emit 2 atomic stores and 1 non-atomic store
+
+// CHECK: func @vector_store_i2_atomic(
+// CHECK-SAME: %[[ARG0:.+]]: vector<7xi2>)
+// CHECK: %[[ALLOC]] = memref.alloc() : memref<6xi8>
+// CHECK: %[[C1:.+]] = arith.constant 1 : index
+// CHECK: %[[CST:.+]] = arith.constant dense<[false, false, false, true]> : vector<4xi1>
+// CHECK: %[[CST0:.+]] = arith.constant dense<0> : vector<4xi2>
+
+// first atomic store
----------------
banach-space wrote:

Where's the store operation?

https://github.com/llvm/llvm-project/pull/115922