[Mlir-commits] [mlir] [mlir][SVE] Add an e2e test for vectorization of linalg.matmul (PR #69592)

[Benjamin Maxwell]

================
@@ -0,0 +1,77 @@
+// RUN: mlir-opt %s -test-transform-dialect-interpreter -test-transform-dialect-erase-schedule \
+// RUN:   -one-shot-bufferize -func-bufferize -cse -canonicalize -convert-vector-to-scf -arm-sve-legalize-vector-storage \
+// RUN:   -convert-vector-to-llvm="enable-arm-sve" -test-lower-to-llvm | \
+// RUN: %mcr_aarch64_cmd -e=entry -entry-point-result=void --march=aarch64 --mattr="+sve" -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils | \
+// RUN: FileCheck %s
+
+func.func @printTestEnd() {
+  %0 = llvm.mlir.addressof @str_sve_end : !llvm.ptr<array<24 x i8>>
+  %1 = llvm.mlir.constant(0 : index) : i64
+  %2 = llvm.getelementptr %0[%1, %1]
+    : (!llvm.ptr<array<24 x i8>>, i64, i64) -> !llvm.ptr<i8>
+  llvm.call @printCString(%2) : (!llvm.ptr<i8>) -> ()
+  return
+}
+
+func.func @entry() {
+  %c1 = arith.constant 1 : index
+  %c2 = arith.constant 2 : index
+  %c4 = arith.constant 4 : index
+  %c0 = arith.constant 0 : index
+  %step = arith.constant 1 : index
+  %c0_f32 = arith.constant 0.0 : f32
+
+  %vscale = vector.vscale
+  %vl_fp = arith.muli %c4, %vscale : index
+  %A_alloc = bufferization.alloc_tensor(%c2, %c1) : tensor<?x?xf32>
+  %B_alloc = bufferization.alloc_tensor(%c1, %vl_fp) : tensor<?x?xf32>
+  %C_alloc = bufferization.alloc_tensor(%c2, %vl_fp) : tensor<?x?xf32>
+
+  %pi = arith.constant  3.14 : f32
+  %A = linalg.fill ins(%pi : f32) outs(%A_alloc : tensor<?x?xf32>) -> tensor<?x?xf32>
+  %B = linalg.fill ins(%pi : f32) outs(%B_alloc : tensor<?x?xf32>) -> tensor<?x?xf32>
+  %C_in = linalg.fill ins(%c0_f32 : f32) outs(%C_alloc : tensor<?x?xf32>) -> tensor<?x?xf32>
+
+  %C_out = linalg.matmul ins(%A, %B: tensor<?x?xf32>, tensor<?x?xf32>) outs(%C_in: tensor<?x?xf32>) -> tensor<?x?xf32>
+
+  // There are at least 4 f32 elements in every SVE vector, i.e. 
+  //    * %vscale is >= 1. 
+  // For implementations with wider vectors, you should see more elements being
+  // printed.
+  // CHECK: {{\[}}[9.8596,   9.8596,   9.8596,   9.8596
+  // CHECK-NEXT: [9.8596,   9.8596,   9.8596,   9.8596
+
+  %xf = tensor.cast %C_out : tensor<?x?xf32> to tensor<*xf32>
+  call @printMemrefF32(%xf) : (tensor<*xf32>) -> ()
+
+  // CHECK: SVE: END OF TEST OUTPUT
+  func.call @printTestEnd() : () -> ()
----------------
MacDue wrote:

Since `vector.print str` is already in your dependencies this can just be: 
```mlir
vector.print str "SVE: END OF TEST OUTPUT"
```
(and `printTestEnd()` and the global can be dropped)

https://github.com/llvm/llvm-project/pull/69592