1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
|
//===- LowerHLFIRIntrinsics.cpp - Transformational intrinsics to FIR ------===//
//
// 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
//
//===----------------------------------------------------------------------===//
#include "flang/Optimizer/Builder/FIRBuilder.h"
#include "flang/Optimizer/Builder/HLFIRTools.h"
#include "flang/Optimizer/Builder/IntrinsicCall.h"
#include "flang/Optimizer/Builder/Todo.h"
#include "flang/Optimizer/Dialect/FIRDialect.h"
#include "flang/Optimizer/Dialect/FIROps.h"
#include "flang/Optimizer/Dialect/FIRType.h"
#include "flang/Optimizer/Dialect/Support/FIRContext.h"
#include "flang/Optimizer/HLFIR/HLFIRDialect.h"
#include "flang/Optimizer/HLFIR/HLFIROps.h"
#include "flang/Optimizer/HLFIR/Passes.h"
#include "mlir/IR/BuiltinDialect.h"
#include "mlir/IR/MLIRContext.h"
#include "mlir/IR/PatternMatch.h"
#include "mlir/Pass/Pass.h"
#include "mlir/Pass/PassManager.h"
#include "mlir/Support/LogicalResult.h"
#include "mlir/Transforms/GreedyPatternRewriteDriver.h"
#include <optional>
namespace hlfir {
#define GEN_PASS_DEF_LOWERHLFIRINTRINSICS
#include "flang/Optimizer/HLFIR/Passes.h.inc"
} // namespace hlfir
namespace {
/// Base class for passes converting transformational intrinsic operations into
/// runtime calls
template <class OP>
class HlfirIntrinsicConversion : public mlir::OpRewritePattern<OP> {
public:
explicit HlfirIntrinsicConversion(mlir::MLIRContext *ctx)
: mlir::OpRewritePattern<OP>{ctx} {
// required for cases where intrinsics are chained together e.g.
// matmul(matmul(a, b), c)
// because converting the inner operation then invalidates the
// outer operation: causing the pattern to apply recursively.
//
// This is safe because we always progress with each iteration. Circular
// applications of operations are not expressible in MLIR because we use
// an SSA form and one must become first. E.g.
// %a = hlfir.matmul %b %d
// %b = hlfir.matmul %a %d
// cannot be written.
// MSVC needs the this->
this->setHasBoundedRewriteRecursion(true);
}
protected:
struct IntrinsicArgument {
mlir::Value val; // allowed to be null if the argument is absent
mlir::Type desiredType;
};
/// Lower the arguments to the intrinsic: adding necessary boxing and
/// conversion to match the signature of the intrinsic in the runtime library.
llvm::SmallVector<fir::ExtendedValue, 3>
lowerArguments(mlir::Operation *op,
const llvm::ArrayRef<IntrinsicArgument> &args,
mlir::PatternRewriter &rewriter,
const fir::IntrinsicArgumentLoweringRules *argLowering) const {
mlir::Location loc = op->getLoc();
fir::FirOpBuilder builder{rewriter, op};
llvm::SmallVector<fir::ExtendedValue, 3> ret;
llvm::SmallVector<std::function<void()>, 2> cleanupFns;
for (size_t i = 0; i < args.size(); ++i) {
mlir::Value arg = args[i].val;
mlir::Type desiredType = args[i].desiredType;
if (!arg) {
ret.emplace_back(fir::getAbsentIntrinsicArgument());
continue;
}
hlfir::Entity entity{arg};
fir::ArgLoweringRule argRules =
fir::lowerIntrinsicArgumentAs(*argLowering, i);
switch (argRules.lowerAs) {
case fir::LowerIntrinsicArgAs::Value: {
if (args[i].desiredType != arg.getType()) {
arg = builder.createConvert(loc, desiredType, arg);
entity = hlfir::Entity{arg};
}
auto [exv, cleanup] = hlfir::convertToValue(loc, builder, entity);
if (cleanup)
cleanupFns.push_back(*cleanup);
ret.emplace_back(exv);
} break;
case fir::LowerIntrinsicArgAs::Addr: {
auto [exv, cleanup] =
hlfir::convertToAddress(loc, builder, entity, desiredType);
if (cleanup)
cleanupFns.push_back(*cleanup);
ret.emplace_back(exv);
} break;
case fir::LowerIntrinsicArgAs::Box: {
auto [box, cleanup] =
hlfir::convertToBox(loc, builder, entity, desiredType);
if (cleanup)
cleanupFns.push_back(*cleanup);
ret.emplace_back(box);
} break;
case fir::LowerIntrinsicArgAs::Inquired: {
if (args[i].desiredType != arg.getType()) {
arg = builder.createConvert(loc, desiredType, arg);
entity = hlfir::Entity{arg};
}
// Place hlfir.expr in memory, and unbox fir.boxchar. Other entities
// are translated to fir::ExtendedValue without transofrmation (notably,
// pointers/allocatable are not dereferenced).
// TODO: once lowering to FIR retires, UBOUND and LBOUND can be
// simplified since the fir.box lowered here are now guarenteed to
// contain the local lower bounds thanks to the hlfir.declare (the extra
// rebox can be removed).
auto [exv, cleanup] =
hlfir::translateToExtendedValue(loc, builder, entity);
if (cleanup)
cleanupFns.push_back(*cleanup);
ret.emplace_back(exv);
} break;
}
}
if (cleanupFns.size()) {
auto oldInsertionPoint = builder.saveInsertionPoint();
builder.setInsertionPointAfter(op);
for (std::function<void()> cleanup : cleanupFns)
cleanup();
builder.restoreInsertionPoint(oldInsertionPoint);
}
return ret;
}
void processReturnValue(mlir::Operation *op,
const fir::ExtendedValue &resultExv, bool mustBeFreed,
fir::FirOpBuilder &builder,
mlir::PatternRewriter &rewriter) const {
mlir::Location loc = op->getLoc();
mlir::Value firBase = fir::getBase(resultExv);
mlir::Type firBaseTy = firBase.getType();
std::optional<hlfir::EntityWithAttributes> resultEntity;
if (fir::isa_trivial(firBaseTy)) {
// Some intrinsics return i1 when the original operation
// produces fir.logical<>, so we may need to cast it.
firBase = builder.createConvert(loc, op->getResult(0).getType(), firBase);
resultEntity = hlfir::EntityWithAttributes{firBase};
} else {
resultEntity =
hlfir::genDeclare(loc, builder, resultExv, ".tmp.intrinsic_result",
fir::FortranVariableFlagsAttr{});
}
if (resultEntity->isVariable()) {
hlfir::AsExprOp asExpr = builder.create<hlfir::AsExprOp>(
loc, *resultEntity, builder.createBool(loc, mustBeFreed));
resultEntity = hlfir::EntityWithAttributes{asExpr.getResult()};
}
mlir::Value base = resultEntity->getBase();
if (!mlir::isa<hlfir::ExprType>(base.getType())) {
for (mlir::Operation *use : op->getResult(0).getUsers()) {
if (mlir::isa<hlfir::DestroyOp>(use))
rewriter.eraseOp(use);
}
}
rewriter.replaceOp(op, base);
}
};
// Given an integer or array of integer type, calculate the Kind parameter from
// the width for use in runtime intrinsic calls.
static unsigned getKindForType(mlir::Type ty) {
mlir::Type eltty = hlfir::getFortranElementType(ty);
unsigned width = mlir::cast<mlir::IntegerType>(eltty).getWidth();
return width / 8;
}
template <class OP>
class HlfirReductionIntrinsicConversion : public HlfirIntrinsicConversion<OP> {
using HlfirIntrinsicConversion<OP>::HlfirIntrinsicConversion;
using IntrinsicArgument =
typename HlfirIntrinsicConversion<OP>::IntrinsicArgument;
using HlfirIntrinsicConversion<OP>::lowerArguments;
using HlfirIntrinsicConversion<OP>::processReturnValue;
protected:
auto buildNumericalArgs(OP operation, mlir::Type i32, mlir::Type logicalType,
mlir::PatternRewriter &rewriter,
std::string opName) const {
llvm::SmallVector<IntrinsicArgument, 3> inArgs;
inArgs.push_back({operation.getArray(), operation.getArray().getType()});
inArgs.push_back({operation.getDim(), i32});
inArgs.push_back({operation.getMask(), logicalType});
auto *argLowering = fir::getIntrinsicArgumentLowering(opName);
return lowerArguments(operation, inArgs, rewriter, argLowering);
};
auto buildMinMaxLocArgs(OP operation, mlir::Type i32, mlir::Type logicalType,
mlir::PatternRewriter &rewriter, std::string opName,
fir::FirOpBuilder builder) const {
llvm::SmallVector<IntrinsicArgument, 3> inArgs;
inArgs.push_back({operation.getArray(), operation.getArray().getType()});
inArgs.push_back({operation.getDim(), i32});
inArgs.push_back({operation.getMask(), logicalType});
mlir::Value kind = builder.createIntegerConstant(
operation->getLoc(), i32, getKindForType(operation.getType()));
inArgs.push_back({kind, i32});
inArgs.push_back({operation.getBack(), i32});
auto *argLowering = fir::getIntrinsicArgumentLowering(opName);
return lowerArguments(operation, inArgs, rewriter, argLowering);
};
auto buildLogicalArgs(OP operation, mlir::Type i32, mlir::Type logicalType,
mlir::PatternRewriter &rewriter,
std::string opName) const {
llvm::SmallVector<IntrinsicArgument, 2> inArgs;
inArgs.push_back({operation.getMask(), logicalType});
inArgs.push_back({operation.getDim(), i32});
auto *argLowering = fir::getIntrinsicArgumentLowering(opName);
return lowerArguments(operation, inArgs, rewriter, argLowering);
};
public:
mlir::LogicalResult
matchAndRewrite(OP operation,
mlir::PatternRewriter &rewriter) const override {
std::string opName;
if constexpr (std::is_same_v<OP, hlfir::SumOp>) {
opName = "sum";
} else if constexpr (std::is_same_v<OP, hlfir::ProductOp>) {
opName = "product";
} else if constexpr (std::is_same_v<OP, hlfir::MaxvalOp>) {
opName = "maxval";
} else if constexpr (std::is_same_v<OP, hlfir::MinvalOp>) {
opName = "minval";
} else if constexpr (std::is_same_v<OP, hlfir::MinlocOp>) {
opName = "minloc";
} else if constexpr (std::is_same_v<OP, hlfir::MaxlocOp>) {
opName = "maxloc";
} else if constexpr (std::is_same_v<OP, hlfir::AnyOp>) {
opName = "any";
} else if constexpr (std::is_same_v<OP, hlfir::AllOp>) {
opName = "all";
} else {
return mlir::failure();
}
fir::FirOpBuilder builder{rewriter, operation.getOperation()};
const mlir::Location &loc = operation->getLoc();
mlir::Type i32 = builder.getI32Type();
mlir::Type logicalType = fir::LogicalType::get(
builder.getContext(), builder.getKindMap().defaultLogicalKind());
llvm::SmallVector<fir::ExtendedValue, 0> args;
if constexpr (std::is_same_v<OP, hlfir::SumOp> ||
std::is_same_v<OP, hlfir::ProductOp> ||
std::is_same_v<OP, hlfir::MaxvalOp> ||
std::is_same_v<OP, hlfir::MinvalOp>) {
args = buildNumericalArgs(operation, i32, logicalType, rewriter, opName);
} else if constexpr (std::is_same_v<OP, hlfir::MinlocOp> ||
std::is_same_v<OP, hlfir::MaxlocOp>) {
args = buildMinMaxLocArgs(operation, i32, logicalType, rewriter, opName,
builder);
} else {
args = buildLogicalArgs(operation, i32, logicalType, rewriter, opName);
}
mlir::Type scalarResultType =
hlfir::getFortranElementType(operation.getType());
auto [resultExv, mustBeFreed] =
fir::genIntrinsicCall(builder, loc, opName, scalarResultType, args);
processReturnValue(operation, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
using SumOpConversion = HlfirReductionIntrinsicConversion<hlfir::SumOp>;
using ProductOpConversion = HlfirReductionIntrinsicConversion<hlfir::ProductOp>;
using MaxvalOpConversion = HlfirReductionIntrinsicConversion<hlfir::MaxvalOp>;
using MinvalOpConversion = HlfirReductionIntrinsicConversion<hlfir::MinvalOp>;
using MinlocOpConversion = HlfirReductionIntrinsicConversion<hlfir::MinlocOp>;
using MaxlocOpConversion = HlfirReductionIntrinsicConversion<hlfir::MaxlocOp>;
using AnyOpConversion = HlfirReductionIntrinsicConversion<hlfir::AnyOp>;
using AllOpConversion = HlfirReductionIntrinsicConversion<hlfir::AllOp>;
struct CountOpConversion : public HlfirIntrinsicConversion<hlfir::CountOp> {
using HlfirIntrinsicConversion<hlfir::CountOp>::HlfirIntrinsicConversion;
mlir::LogicalResult
matchAndRewrite(hlfir::CountOp count,
mlir::PatternRewriter &rewriter) const override {
fir::FirOpBuilder builder{rewriter, count.getOperation()};
const mlir::Location &loc = count->getLoc();
mlir::Type i32 = builder.getI32Type();
mlir::Type logicalType = fir::LogicalType::get(
builder.getContext(), builder.getKindMap().defaultLogicalKind());
llvm::SmallVector<IntrinsicArgument, 3> inArgs;
inArgs.push_back({count.getMask(), logicalType});
inArgs.push_back({count.getDim(), i32});
mlir::Value kind = builder.createIntegerConstant(
count->getLoc(), i32, getKindForType(count.getType()));
inArgs.push_back({kind, i32});
auto *argLowering = fir::getIntrinsicArgumentLowering("count");
llvm::SmallVector<fir::ExtendedValue, 3> args =
lowerArguments(count, inArgs, rewriter, argLowering);
mlir::Type scalarResultType = hlfir::getFortranElementType(count.getType());
auto [resultExv, mustBeFreed] =
fir::genIntrinsicCall(builder, loc, "count", scalarResultType, args);
processReturnValue(count, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
struct MatmulOpConversion : public HlfirIntrinsicConversion<hlfir::MatmulOp> {
using HlfirIntrinsicConversion<hlfir::MatmulOp>::HlfirIntrinsicConversion;
mlir::LogicalResult
matchAndRewrite(hlfir::MatmulOp matmul,
mlir::PatternRewriter &rewriter) const override {
fir::FirOpBuilder builder{rewriter, matmul.getOperation()};
const mlir::Location &loc = matmul->getLoc();
mlir::Value lhs = matmul.getLhs();
mlir::Value rhs = matmul.getRhs();
llvm::SmallVector<IntrinsicArgument, 2> inArgs;
inArgs.push_back({lhs, lhs.getType()});
inArgs.push_back({rhs, rhs.getType()});
auto *argLowering = fir::getIntrinsicArgumentLowering("matmul");
llvm::SmallVector<fir::ExtendedValue, 2> args =
lowerArguments(matmul, inArgs, rewriter, argLowering);
mlir::Type scalarResultType =
hlfir::getFortranElementType(matmul.getType());
auto [resultExv, mustBeFreed] =
fir::genIntrinsicCall(builder, loc, "matmul", scalarResultType, args);
processReturnValue(matmul, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
struct DotProductOpConversion
: public HlfirIntrinsicConversion<hlfir::DotProductOp> {
using HlfirIntrinsicConversion<hlfir::DotProductOp>::HlfirIntrinsicConversion;
mlir::LogicalResult
matchAndRewrite(hlfir::DotProductOp dotProduct,
mlir::PatternRewriter &rewriter) const override {
fir::FirOpBuilder builder{rewriter, dotProduct.getOperation()};
const mlir::Location &loc = dotProduct->getLoc();
mlir::Value lhs = dotProduct.getLhs();
mlir::Value rhs = dotProduct.getRhs();
llvm::SmallVector<IntrinsicArgument, 2> inArgs;
inArgs.push_back({lhs, lhs.getType()});
inArgs.push_back({rhs, rhs.getType()});
auto *argLowering = fir::getIntrinsicArgumentLowering("dot_product");
llvm::SmallVector<fir::ExtendedValue, 2> args =
lowerArguments(dotProduct, inArgs, rewriter, argLowering);
mlir::Type scalarResultType =
hlfir::getFortranElementType(dotProduct.getType());
auto [resultExv, mustBeFreed] = fir::genIntrinsicCall(
builder, loc, "dot_product", scalarResultType, args);
processReturnValue(dotProduct, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
class TransposeOpConversion
: public HlfirIntrinsicConversion<hlfir::TransposeOp> {
using HlfirIntrinsicConversion<hlfir::TransposeOp>::HlfirIntrinsicConversion;
mlir::LogicalResult
matchAndRewrite(hlfir::TransposeOp transpose,
mlir::PatternRewriter &rewriter) const override {
fir::FirOpBuilder builder{rewriter, transpose.getOperation()};
const mlir::Location &loc = transpose->getLoc();
mlir::Value arg = transpose.getArray();
llvm::SmallVector<IntrinsicArgument, 1> inArgs;
inArgs.push_back({arg, arg.getType()});
auto *argLowering = fir::getIntrinsicArgumentLowering("transpose");
llvm::SmallVector<fir::ExtendedValue, 1> args =
lowerArguments(transpose, inArgs, rewriter, argLowering);
mlir::Type scalarResultType =
hlfir::getFortranElementType(transpose.getType());
auto [resultExv, mustBeFreed] = fir::genIntrinsicCall(
builder, loc, "transpose", scalarResultType, args);
processReturnValue(transpose, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
struct MatmulTransposeOpConversion
: public HlfirIntrinsicConversion<hlfir::MatmulTransposeOp> {
using HlfirIntrinsicConversion<
hlfir::MatmulTransposeOp>::HlfirIntrinsicConversion;
mlir::LogicalResult
matchAndRewrite(hlfir::MatmulTransposeOp multranspose,
mlir::PatternRewriter &rewriter) const override {
fir::FirOpBuilder builder{rewriter, multranspose.getOperation()};
const mlir::Location &loc = multranspose->getLoc();
mlir::Value lhs = multranspose.getLhs();
mlir::Value rhs = multranspose.getRhs();
llvm::SmallVector<IntrinsicArgument, 2> inArgs;
inArgs.push_back({lhs, lhs.getType()});
inArgs.push_back({rhs, rhs.getType()});
auto *argLowering = fir::getIntrinsicArgumentLowering("matmul");
llvm::SmallVector<fir::ExtendedValue, 2> args =
lowerArguments(multranspose, inArgs, rewriter, argLowering);
mlir::Type scalarResultType =
hlfir::getFortranElementType(multranspose.getType());
auto [resultExv, mustBeFreed] = fir::genIntrinsicCall(
builder, loc, "matmul_transpose", scalarResultType, args);
processReturnValue(multranspose, resultExv, mustBeFreed, builder, rewriter);
return mlir::success();
}
};
class LowerHLFIRIntrinsics
: public hlfir::impl::LowerHLFIRIntrinsicsBase<LowerHLFIRIntrinsics> {
public:
void runOnOperation() override {
mlir::ModuleOp module = this->getOperation();
mlir::MLIRContext *context = &getContext();
mlir::RewritePatternSet patterns(context);
patterns
.insert<MatmulOpConversion, MatmulTransposeOpConversion,
AllOpConversion, AnyOpConversion, SumOpConversion,
ProductOpConversion, TransposeOpConversion, CountOpConversion,
DotProductOpConversion, MaxvalOpConversion, MinvalOpConversion,
MinlocOpConversion, MaxlocOpConversion>(context);
// While conceptually this pass is performing dialect conversion, we use
// pattern rewrites here instead of dialect conversion because this pass
// looses array bounds from some of the expressions e.g.
// !hlfir.expr<2xi32> -> !hlfir.expr<?xi32>
// MLIR thinks this is a different type so dialect conversion fails.
// Pattern rewriting only requires that the resulting IR is still valid
mlir::GreedyRewriteConfig config;
// Prevent the pattern driver from merging blocks
config.enableRegionSimplification = false;
if (mlir::failed(mlir::applyPatternsAndFoldGreedily(
module, std::move(patterns), config))) {
mlir::emitError(mlir::UnknownLoc::get(context),
"failure in HLFIR intrinsic lowering");
signalPassFailure();
}
}
};
} // namespace
|