[Mlir-commits] [mlir] [mlir][SVE] Add an e2e test for vector.contract (PR #69845)

[Benjamin Maxwell]

================
@@ -0,0 +1,137 @@
+// DEFINE: %{compile} = mlir-opt %s  -test-transform-dialect-interpreter -test-transform-dialect-erase-schedule\
+// DEFINE:    -cse -canonicalize -convert-vector-to-scf -arm-sve-legalize-vector-storage\
+// DEFINE:    -convert-vector-to-llvm="enable-arm-sve" -test-lower-to-llvm
+// DEFINE: %{entry} =
+// DEFINE: %{run} = %mcr_aarch64_cmd -e=%{entry} -entry-point-result=void --march=aarch64 --mattr="+sve" -shared-libs=%mlir_lib_dir/libmlir_c_runner_utils%shlibext
+
+// REDEFINE: %{entry} = entry_i32
+// RUN: %{compile} | %{run} | FileCheck %s --check-prefix=I32
+
+// REDEFINE: %{entry} = entry_f32
+// RUN: %{compile} | %{run} | FileCheck %s --check-prefix=F32
+
+#matmat_accesses = [
+  affine_map<(i, j, k) -> (i, k)>,
+  affine_map<(i, j, k) -> (k, j)>,
+  affine_map<(i, j, k) -> (i, j)>
+]
+#matmat_trait = {
+  indexing_maps = #matmat_accesses,
+  iterator_types = ["parallel", "parallel", "reduction"]
+}
+
+func.func @entry_i32() {
+  %vscale = vector.vscale
+
+  %c0 = arith.constant 0 : index
+  %c2 = arith.constant 2 : index
+  %c3 = arith.constant 3 : index
+  %c5 = arith.constant 5 : index
+  %n_rows = arith.muli %vscale, %c2 : index
+
+  %cst = arith.constant 0: i32
+  %i32_123 = arith.constant 123 : i32
+  %i32_314 = arith.constant 314 : i32
+
+  // Allocate and initialize matrix A
+  %A_alloc = memref.alloca() : memref<3x5xi32>
+  linalg.fill ins(%i32_123 : i32) outs(%A_alloc :memref<3x5xi32>)
+  %mask_a = vector.create_mask %c3, %c5 : vector<3x5xi1>
+  %vector_a = vector.transfer_read %A_alloc[%c0, %c0], %cst, %mask_a {in_bounds = [true, true]} : memref<3x5xi32>, vector<3x5xi32>
----------------
MacDue wrote:

Are these memrefs/loads needed?
```suggestion
  // Setup vector A:
  %vector_a = arith.constant : dense<123> : vector<3x5xi32>
```
(same comment elsewhere) 

https://github.com/llvm/llvm-project/pull/69845