Add a pass that specializes parallel loops for easier unrolling and vectorization

This matches loops with a affine.min upper bound, limiting the trip
count to a constant, and rewrites them into two loops, one with constant
upper bound and one with variable upper bound. The assumption is that
the constant upper bound loop will be unrolled and vectorized, which is
preferable if this is the hot path.

Differential Revision: https://reviews.llvm.org/D75240

diff --git a/mlir/include/mlir/Dialect/LoopOps/Passes.h b/mlir/include/mlir/Dialect/LoopOps/Passes.h
index b33b5ae..54c0fb4 100644 (file)
--- a/mlir/include/mlir/Dialect/LoopOps/Passes.h
+++ b/mlir/include/mlir/Dialect/LoopOps/Passes.h
@@ -23,6 +23,10 @@ class Pass;
 /// Creates a loop fusion pass which fuses parallel loops.
 std::unique_ptr<Pass> createParallelLoopFusionPass();
 
+/// Creates a pass that specializes parallel loop for unrolling and
+/// vectorization.
+std::unique_ptr<Pass> createParallelLoopSpecializationPass();
+
 /// Creates a pass which tiles innermost parallel loops.
 std::unique_ptr<Pass>
 createParallelLoopTilingPass(llvm::ArrayRef<int64_t> tileSize = {});

diff --git a/mlir/include/mlir/InitAllPasses.h b/mlir/include/mlir/InitAllPasses.h
index a1edc82..2b21ccb 100644 (file)
--- a/mlir/include/mlir/InitAllPasses.h
+++ b/mlir/include/mlir/InitAllPasses.h
@@ -109,6 +109,7 @@ inline void registerAllPasses() {
 
   // LoopOps
   createParallelLoopFusionPass();
+  createParallelLoopSpecializationPass();
   createParallelLoopTilingPass();
 
   // QuantOps

diff --git a/mlir/lib/Dialect/LoopOps/Transforms/CMakeLists.txt b/mlir/lib/Dialect/LoopOps/Transforms/CMakeLists.txt
index 9dbd5a3..7f5b16f 100644 (file)
--- a/mlir/lib/Dialect/LoopOps/Transforms/CMakeLists.txt
+++ b/mlir/lib/Dialect/LoopOps/Transforms/CMakeLists.txt
@@ -1,5 +1,6 @@
 add_llvm_library(MLIRLoopOpsTransforms
   ParallelLoopFusion.cpp
+  ParallelLoopSpecialization.cpp
   ParallelLoopTiling.cpp
 
   ADDITIONAL_HEADER_DIRS

diff --git a/mlir/lib/Dialect/LoopOps/Transforms/ParallelLoopSpecialization.cpp b/mlir/lib/Dialect/LoopOps/Transforms/ParallelLoopSpecialization.cpp
new file mode 100644 (file)
index 0000000..8cb49f3
--- /dev/null
+++ b/mlir/lib/Dialect/LoopOps/Transforms/ParallelLoopSpecialization.cpp
@@ -0,0 +1,76 @@
+//===- ParallelLoopSpecialization.cpp - loop.parallel specializeation -----===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+//
+// Specializes parallel loops for easier unrolling and vectorization.
+//
+//===----------------------------------------------------------------------===//
+
+#include "mlir/Dialect/AffineOps/AffineOps.h"
+#include "mlir/Dialect/LoopOps/LoopOps.h"
+#include "mlir/Dialect/LoopOps/Passes.h"
+#include "mlir/Dialect/StandardOps/IR/Ops.h"
+#include "mlir/IR/AffineExpr.h"
+#include "mlir/IR/BlockAndValueMapping.h"
+#include "mlir/Pass/Pass.h"
+
+using namespace mlir;
+using loop::ParallelOp;
+
+/// Rewrite a loop with bounds defined by an affine.min with a constant into 2
+/// loops after checking if the bounds are equal to that constant. This is
+/// beneficial if the loop will almost always have the constant bound and that
+/// version can be fully unrolled and vectorized.
+static void specializeLoopForUnrolling(ParallelOp op) {
+  SmallVector<int64_t, 2> constantIndices;
+  constantIndices.reserve(op.upperBound().size());
+  for (auto bound : op.upperBound()) {
+    auto minOp = dyn_cast_or_null<AffineMinOp>(bound.getDefiningOp());
+    if (!minOp)
+      return;
+    int64_t minConstant = std::numeric_limits<int64_t>::max();
+    for (auto expr : minOp.map().getResults()) {
+      if (auto constantIndex = expr.dyn_cast<AffineConstantExpr>())
+        minConstant = std::min(minConstant, constantIndex.getValue());
+    }
+    if (minConstant == std::numeric_limits<int64_t>::max())
+      return;
+    constantIndices.push_back(minConstant);
+  }
+
+  OpBuilder b(op);
+  BlockAndValueMapping map;
+  Value cond;
+  for (auto bound : llvm::zip(op.upperBound(), constantIndices)) {
+    Value constant = b.create<ConstantIndexOp>(op.getLoc(), std::get<1>(bound));
+    Value cmp = b.create<CmpIOp>(op.getLoc(), CmpIPredicate::eq,
+                                 std::get<0>(bound), constant);
+    cond = cond ? b.create<AndOp>(op.getLoc(), cond, cmp) : cmp;
+    map.map(std::get<0>(bound), constant);
+  }
+  auto ifOp = b.create<loop::IfOp>(op.getLoc(), cond, /*withElseRegion=*/true);
+  ifOp.getThenBodyBuilder().clone(*op.getOperation(), map);
+  ifOp.getElseBodyBuilder().clone(*op.getOperation());
+  op.erase();
+}
+
+namespace {
+struct ParallelLoopSpecialization
+    : public FunctionPass<ParallelLoopSpecialization> {
+  void runOnFunction() override {
+    getFunction().walk([](ParallelOp op) { specializeLoopForUnrolling(op); });
+  }
+};
+} // namespace
+
+std::unique_ptr<Pass> mlir::createParallelLoopSpecializationPass() {
+  return std::make_unique<ParallelLoopSpecialization>();
+}
+
+static PassRegistration<ParallelLoopSpecialization>
+    pass("parallel-loop-specialization",
+         "Specialize parallel loops for vectorization.");

diff --git a/mlir/test/Dialect/Loops/parallel-loop-specialization.mlir b/mlir/test/Dialect/Loops/parallel-loop-specialization.mlir
new file mode 100644 (file)
index 0000000..ab736c9
--- /dev/null
+++ b/mlir/test/Dialect/Loops/parallel-loop-specialization.mlir
@@ -0,0 +1,46 @@
+// RUN: mlir-opt %s -parallel-loop-specialization -split-input-file | FileCheck %s --dump-input-on-failure
+
+#map0 = affine_map<()[s0, s1] -> (1024, s0 - s1)>
+#map1 = affine_map<()[s0, s1] -> (64, s0 - s1)>
+
+func @parallel_loop(%outer_i0: index, %outer_i1: index, %A: memref<?x?xf32>, %B: memref<?x?xf32>,
+                    %C: memref<?x?xf32>, %result: memref<?x?xf32>) {
+  %c0 = constant 0 : index
+  %c1 = constant 1 : index
+  %d0 = dim %A, 0 : memref<?x?xf32>
+  %d1 = dim %A, 1 : memref<?x?xf32>
+  %b0 = affine.min #map0()[%d0, %outer_i0]
+  %b1 = affine.min #map1()[%d1, %outer_i1]
+  loop.parallel (%i0, %i1) = (%c0, %c0) to (%b0, %b1) step (%c1, %c1) {
+    %B_elem = load %B[%i0, %i1] : memref<?x?xf32>
+    %C_elem = load %C[%i0, %i1] : memref<?x?xf32>
+    %sum_elem = addf %B_elem, %C_elem : f32
+    store %sum_elem, %result[%i0, %i1] : memref<?x?xf32>
+  }
+  return
+}
+
+// CHECK-LABEL:   func @parallel_loop(
+// CHECK-SAME:                        [[VAL_0:%.*]]: index, [[VAL_1:%.*]]: index, [[VAL_2:%.*]]: memref<?x?xf32>, [[VAL_3:%.*]]: memref<?x?xf32>, [[VAL_4:%.*]]: memref<?x?xf32>, [[VAL_5:%.*]]: memref<?x?xf32>) {
+// CHECK:           [[VAL_6:%.*]] = constant 0 : index
+// CHECK:           [[VAL_7:%.*]] = constant 1 : index
+// CHECK:           [[VAL_8:%.*]] = dim [[VAL_2]], 0 : memref<?x?xf32>
+// CHECK:           [[VAL_9:%.*]] = dim [[VAL_2]], 1 : memref<?x?xf32>
+// CHECK:           [[VAL_10:%.*]] = affine.min #map0(){{\[}}[[VAL_8]], [[VAL_0]]]
+// CHECK:           [[VAL_11:%.*]] = affine.min #map1(){{\[}}[[VAL_9]], [[VAL_1]]]
+// CHECK:           [[VAL_12:%.*]] = constant 1024 : index
+// CHECK:           [[VAL_13:%.*]] = cmpi "eq", [[VAL_10]], [[VAL_12]] : index
+// CHECK:           [[VAL_14:%.*]] = constant 64 : index
+// CHECK:           [[VAL_15:%.*]] = cmpi "eq", [[VAL_11]], [[VAL_14]] : index
+// CHECK:           [[VAL_16:%.*]] = and [[VAL_13]], [[VAL_15]] : i1
+// CHECK:           loop.if [[VAL_16]] {
+// CHECK:             loop.parallel ([[VAL_17:%.*]], [[VAL_18:%.*]]) = ([[VAL_6]], [[VAL_6]]) to ([[VAL_12]], [[VAL_14]]) step ([[VAL_7]], [[VAL_7]]) {
+// CHECK:               store
+// CHECK:             }
+// CHECK:           } else {
+// CHECK:             loop.parallel ([[VAL_22:%.*]], [[VAL_23:%.*]]) = ([[VAL_6]], [[VAL_6]]) to ([[VAL_10]], [[VAL_11]]) step ([[VAL_7]], [[VAL_7]]) {
+// CHECK:               store
+// CHECK:             }
+// CHECK:           }
+// CHECK:           return
+// CHECK:         }