[llvm] [mlir] [mlir][ArmNeon] Implements LowerVectorToArmNeon Pattern for SMMLA (PR #81895)

[Diego Caballero via llvm-commits]

================
@@ -0,0 +1,153 @@
+//===- LowerVectorToArmNeon.cpp - Lower 'arm_neon.intr.smmla' ops
+//-----------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file implements lowering patterns from vector.contract to
+// arm_neon.intr.smmla
+//
+//===---
+
+#include "mlir/Dialect/Arith/IR/Arith.h"
+#include "mlir/Dialect/ArmNeon/ArmNeonDialect.h"
+#include "mlir/Dialect/ArmNeon/Transforms.h"
+#include "mlir/Dialect/Func/IR/FuncOps.h"
+#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
+#include "mlir/Dialect/Vector/IR/VectorOps.h"
+#include "mlir/IR/PatternMatch.h"
+#include "mlir/Support/LogicalResult.h"
+#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
+
+#define DEBUG_TYPE "arm-neon-lower-vector"
+
+using namespace mlir;
+using namespace mlir::arm_neon;
+
+namespace {
+
+// Return the shaped type with new element type.
+static Type matchContainerType(Type element, Type container) {
+  if (auto shapedTy = dyn_cast<ShapedType>(container))
+    return shapedTy.clone(element);
+
+  return element;
+}
+
+// Lowering from vector::contractOp directly to the arm neon
+// intrinsic.
+class LowerVectorToArmNeonPattern
----------------
dcaballe wrote:

Probably good to make this name more explicit... Something like `LowerContractionToSMMLAPattern`?

https://github.com/llvm/llvm-project/pull/81895