[mlir][SVE] Add an e2e test for vectorization of linalg.matmul

[banach-space]

Adds an end-to-end test for scalable vectorization of linalg.matmul.

[llvmbot]

@llvm/pr-subscribers-mlir-sve
@llvm/pr-subscribers-mlir

@llvm/pr-subscribers-mlir-linalg

Author: Andrzej Warzyński (banach-space)

Changes

Adds an end-to-end test for scalable vectorization of linalg.matmul.

---

Full diff: https://github.com/llvm/llvm-project/pull/70372.diff

1 Files Affected:

• (added) mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/matmul.mlir (+68)

diff --git a/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/matmul.mlir b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/matmul.mlir
new file mode 100644
index 000000000000000..2024da2a585d99f
--- /dev/null
+++ b/mlir/test/Integration/Dialect/Linalg/CPU/ArmSVE/matmul.mlir
@@ -0,0 +1,68 @@
+// 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=matmul_f32 -entry-point-result=void --march=aarch64 --mattr="+sve" -shared-libs=%mlir_runner_utils,%mlir_c_runner_utils | \
+// RUN: FileCheck %s
+
+func.func @matmul_f32() {
+  // Matrix dimensions
+  %K = arith.constant 3 : index
+  %M = arith.constant 5 : index
+  %N = arith.constant 15 : index
+  %c0_f32 = arith.constant 0.0 : f32
+
+  // Allocate the matrices
+  %A_alloc = bufferization.alloc_tensor(%M, %K) : tensor<?x?xf32>
+  %B_alloc = bufferization.alloc_tensor(%K, %N) : tensor<?x?xf32>
+  %C_alloc = bufferization.alloc_tensor(%M, %N) : tensor<?x?xf32>
+
+  // Initialise the matrices
+  %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>
+
+  // Matmul
+  %C_out = linalg.matmul ins(%A, %B: tensor<?x?xf32>, tensor<?x?xf32>) outs(%C_in: tensor<?x?xf32>) -> tensor<?x?xf32>
+
+  // Print and verify the output
+  // CHECK-LABEL: SVE: START OF TEST OUTPUT
+  vector.print str "SVE: START OF TEST OUTPUT"
+
+  // CHECK-NEXT: Unranked Memref {{.*}} rank = 2 offset = 0 sizes = [5, 15] strides = [15, 1] data =
+  // CHECK-COUNT-5: [29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788, 29.5788]
+  %xf = tensor.cast %C_out : tensor<?x?xf32> to tensor<*xf32>
+  call @printMemrefF32(%xf) : (tensor<*xf32>) -> ()
+
+  // CHECK-NEXT: SVE: END OF TEST OUTPUT
+  vector.print str "SVE: END OF TEST OUTPUT"
+
+  return
+}
+
+transform.sequence failures(propagate) {
+^bb1(%module_op: !transform.any_op):
+  // Step 1: Tile
+  %matmul = transform.structured.match ops{["linalg.matmul"]} in %module_op : (!transform.any_op) -> !transform.any_op
+  %func_op = get_parent_op %matmul : (!transform.any_op) -> !transform.op<"func.func">
+  %module_with_tiled_loops, %loops:3 = transform.structured.tile_using_for %matmul [2, [4], 1] : (!transform.any_op) -> (!transform.any_op, !transform.any_op, !transform.any_op, !transform.any_op)
+
+  // Step 2: Vectorize
+  %tiled_matmul = transform.structured.match ops{["linalg.matmul"]} in %module_with_tiled_loops : (!transform.any_op) -> !transform.any_op
+  transform.structured.vectorize %tiled_matmul vector_sizes [2, [4], 1] : !transform.any_op
+
+  // Step 3: Lower vector.multi_reduction to vector.contract (+ some helpful patterns)
+  transform.apply_patterns to %func_op {
+    transform.apply_patterns.vector.reduction_to_contract
+    transform.apply_patterns.vector.transfer_permutation_patterns
+    transform.apply_patterns.vector.lower_masked_transfers
+  } : !transform.op<"func.func">
+
+  // Step 4: Lower vector.contract to vector.fma
+  transform.apply_patterns to %func_op {
+    transform.apply_patterns.vector.lower_contraction lowering_strategy = "outerproduct"
+    transform.apply_patterns.vector.lower_outerproduct
+  } : !transform.op<"func.func">
+}
+
+func.func private @printMemrefF32(%ptr : tensor<*xf32>)

[banach-space]

@MacDue , rather than fixing #69592, I decided to upgrade it a bit (the bot failure made me realize that it should not depend on the SVE implementation). Hence a new PR. Apologies for the noise!

[MacDue]

Makes sense to me, keeping the matrices a fixed-size seems like a more realistic test 👍