|
| 1 | +//=- VectorEmulateMaskedLoadStore.cpp - Emulate 'vector.maskedload/store' op =// |
| 2 | +// |
| 3 | +// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | +// See https://llvm.org/LICENSE.txt for license information. |
| 5 | +// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | +// |
| 7 | +//===----------------------------------------------------------------------===// |
| 8 | +// |
| 9 | +// This file implements target-independent rewrites and utilities to emulate the |
| 10 | +// 'vector.maskedload' and 'vector.maskedstore' operation. |
| 11 | +// |
| 12 | +//===----------------------------------------------------------------------===// |
| 13 | + |
| 14 | +#include "mlir/Dialect/MemRef/IR/MemRef.h" |
| 15 | +#include "mlir/Dialect/SCF/IR/SCF.h" |
| 16 | +#include "mlir/Dialect/Vector/Transforms/LoweringPatterns.h" |
| 17 | + |
| 18 | +using namespace mlir; |
| 19 | + |
| 20 | +namespace { |
| 21 | + |
| 22 | +/// Convert vector.maskedload |
| 23 | +/// |
| 24 | +/// Before: |
| 25 | +/// |
| 26 | +/// vector.maskedload %base[%idx_0, %idx_1], %mask, %pass_thru |
| 27 | +/// |
| 28 | +/// After: |
| 29 | +/// |
| 30 | +/// %ivalue = %pass_thru |
| 31 | +/// %m = vector.extract %mask[0] |
| 32 | +/// %result0 = scf.if %m { |
| 33 | +/// %v = memref.load %base[%idx_0, %idx_1] |
| 34 | +/// %combined = vector.insert %v, %ivalue[0] |
| 35 | +/// scf.yield %combined |
| 36 | +/// } else { |
| 37 | +/// scf.yield %ivalue |
| 38 | +/// } |
| 39 | +/// %m = vector.extract %mask[1] |
| 40 | +/// %result1 = scf.if %m { |
| 41 | +/// %v = memref.load %base[%idx_0, %idx_1 + 1] |
| 42 | +/// %combined = vector.insert %v, %result0[1] |
| 43 | +/// scf.yield %combined |
| 44 | +/// } else { |
| 45 | +/// scf.yield %result0 |
| 46 | +/// } |
| 47 | +/// ... |
| 48 | +/// |
| 49 | +struct VectorMaskedLoadOpConverter final |
| 50 | + : OpRewritePattern<vector::MaskedLoadOp> { |
| 51 | + using OpRewritePattern::OpRewritePattern; |
| 52 | + |
| 53 | + LogicalResult matchAndRewrite(vector::MaskedLoadOp maskedLoadOp, |
| 54 | + PatternRewriter &rewriter) const override { |
| 55 | + VectorType maskVType = maskedLoadOp.getMaskVectorType(); |
| 56 | + if (maskVType.getShape().size() != 1) |
| 57 | + return rewriter.notifyMatchFailure( |
| 58 | + maskedLoadOp, "expected vector.maskedstore with 1-D mask"); |
| 59 | + |
| 60 | + Location loc = maskedLoadOp.getLoc(); |
| 61 | + int64_t maskLength = maskVType.getShape()[0]; |
| 62 | + |
| 63 | + Type indexType = rewriter.getIndexType(); |
| 64 | + Value mask = maskedLoadOp.getMask(); |
| 65 | + Value base = maskedLoadOp.getBase(); |
| 66 | + Value iValue = maskedLoadOp.getPassThru(); |
| 67 | + auto indices = llvm::to_vector_of<Value>(maskedLoadOp.getIndices()); |
| 68 | + Value one = rewriter. create<arith::ConstantOp>( |
| 69 | + loc, indexType, IntegerAttr::get(indexType, 1)); |
| 70 | + for (int64_t i = 0; i < maskLength; ++i) { |
| 71 | + auto maskBit = rewriter.create<vector::ExtractOp>(loc, mask, i); |
| 72 | + |
| 73 | + auto ifOp = rewriter.create<scf::IfOp>( |
| 74 | + loc, maskBit, |
| 75 | + [&](OpBuilder &builder, Location loc) { |
| 76 | + auto loadedValue = |
| 77 | + builder.create<memref::LoadOp>(loc, base, indices); |
| 78 | + auto combinedValue = |
| 79 | + builder.create<vector::InsertOp>(loc, loadedValue, iValue, i); |
| 80 | + builder.create<scf::YieldOp>(loc, combinedValue.getResult()); |
| 81 | + }, |
| 82 | + [&](OpBuilder &builder, Location loc) { |
| 83 | + builder.create<scf::YieldOp>(loc, iValue); |
| 84 | + }); |
| 85 | + iValue = ifOp.getResult(0); |
| 86 | + |
| 87 | + indices. back() = rewriter. create<arith::AddIOp>(loc, indices. back(), one); |
| 88 | + } |
| 89 | + |
| 90 | + rewriter.replaceOp(maskedLoadOp, iValue); |
| 91 | + |
| 92 | + return success(); |
| 93 | + } |
| 94 | +}; |
| 95 | + |
| 96 | +/// Convert vector.maskedstore |
| 97 | +/// |
| 98 | +/// Before: |
| 99 | +/// |
| 100 | +/// vector.maskedstore %base[%idx_0, %idx_1], %mask, %value |
| 101 | +/// |
| 102 | +/// After: |
| 103 | +/// |
| 104 | +/// %m = vector.extract %mask[0] |
| 105 | +/// scf.if %m { |
| 106 | +/// %extracted = vector.extract %value[0] |
| 107 | +/// memref.store %extracted, %base[%idx_0, %idx_1] |
| 108 | +/// } |
| 109 | +/// %m = vector.extract %mask[1] |
| 110 | +/// scf.if %m { |
| 111 | +/// %extracted = vector.extract %value[1] |
| 112 | +/// memref.store %extracted, %base[%idx_0, %idx_1 + 1] |
| 113 | +/// } |
| 114 | +/// ... |
| 115 | +/// |
| 116 | +struct VectorMaskedStoreOpConverter final |
| 117 | + : OpRewritePattern<vector::MaskedStoreOp> { |
| 118 | + using OpRewritePattern::OpRewritePattern; |
| 119 | + |
| 120 | + LogicalResult matchAndRewrite(vector::MaskedStoreOp maskedStoreOp, |
| 121 | + PatternRewriter &rewriter) const override { |
| 122 | + VectorType maskVType = maskedStoreOp.getMaskVectorType(); |
| 123 | + if (maskVType.getShape().size() != 1) |
| 124 | + return rewriter.notifyMatchFailure( |
| 125 | + maskedStoreOp, "expected vector.maskedstore with 1-D mask"); |
| 126 | + |
| 127 | + Location loc = maskedStoreOp.getLoc(); |
| 128 | + int64_t maskLength = maskVType.getShape()[0]; |
| 129 | + |
| 130 | + Type indexType = rewriter.getIndexType(); |
| 131 | + Value mask = maskedStoreOp.getMask(); |
| 132 | + Value base = maskedStoreOp.getBase(); |
| 133 | + Value value = maskedStoreOp.getValueToStore(); |
| 134 | + auto indices = llvm::to_vector_of<Value>(maskedStoreOp.getIndices()); |
| 135 | + Value one = rewriter. create<arith::ConstantOp>( |
| 136 | + loc, indexType, IntegerAttr::get(indexType, 1)); |
| 137 | + for (int64_t i = 0; i < maskLength; ++i) { |
| 138 | + auto maskBit = rewriter.create<vector::ExtractOp>(loc, mask, i); |
| 139 | + |
| 140 | + auto ifOp = rewriter.create<scf::IfOp>(loc, maskBit, /*else=*/false); |
| 141 | + rewriter.setInsertionPointToStart(&ifOp.getThenRegion().front()); |
| 142 | + auto extractedValue = rewriter.create<vector::ExtractOp>(loc, value, i); |
| 143 | + rewriter.create<memref::StoreOp>(loc, extractedValue, base, indices); |
| 144 | + |
| 145 | + rewriter.setInsertionPointAfter(ifOp); |
| 146 | + indices. back() = rewriter. create<arith::AddIOp>(loc, indices. back(), one); |
| 147 | + } |
| 148 | + |
| 149 | + rewriter.eraseOp(maskedStoreOp); |
| 150 | + |
| 151 | + return success(); |
| 152 | + } |
| 153 | +}; |
| 154 | + |
| 155 | +} // namespace |
| 156 | + |
| 157 | +void mlir::vector::populateVectorMaskedLoadStoreEmulationPatterns( |
| 158 | + RewritePatternSet &patterns, PatternBenefit benefit) { |
| 159 | + patterns.add<VectorMaskedLoadOpConverter, VectorMaskedStoreOpConverter>( |
| 160 | + patterns.getContext(), benefit); |
| 161 | +} |
0 commit comments