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
|
//===- KernelOutlining.cpp - Implementation of GPU kernel outlining -------===//
//
// Part of the MLIR 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 the GPU dialect kernel outlining pass.
//
//===----------------------------------------------------------------------===//
#include "mlir/Dialect/GPU/GPUDialect.h"
#include "mlir/Dialect/GPU/Passes.h"
#include "mlir/Dialect/StandardOps/Ops.h"
#include "mlir/IR/BlockAndValueMapping.h"
#include "mlir/IR/Builders.h"
#include "mlir/IR/SymbolTable.h"
#include "mlir/Pass/Pass.h"
using namespace mlir;
template <typename OpTy>
static void createForAllDimensions(OpBuilder &builder, Location loc,
SmallVectorImpl<ValuePtr> &values) {
for (StringRef dim : {"x", "y", "z"}) {
ValuePtr v = builder.create<OpTy>(loc, builder.getIndexType(),
builder.getStringAttr(dim));
values.push_back(v);
}
}
// Add operations generating block/thread ids and grid/block dimensions at the
// beginning of the `body` region and replace uses of the respective function
// arguments.
static void injectGpuIndexOperations(Location loc, Region &body) {
OpBuilder builder(loc->getContext());
Block &firstBlock = body.front();
builder.setInsertionPointToStart(&firstBlock);
SmallVector<ValuePtr, 12> indexOps;
createForAllDimensions<gpu::BlockIdOp>(builder, loc, indexOps);
createForAllDimensions<gpu::ThreadIdOp>(builder, loc, indexOps);
createForAllDimensions<gpu::GridDimOp>(builder, loc, indexOps);
createForAllDimensions<gpu::BlockDimOp>(builder, loc, indexOps);
// Replace the leading 12 function args with the respective thread/block index
// operations. Iterate backwards since args are erased and indices change.
for (int i = 11; i >= 0; --i) {
firstBlock.getArgument(i)->replaceAllUsesWith(indexOps[i]);
firstBlock.eraseArgument(i);
}
}
static bool isInliningBeneficiary(Operation *op) {
return isa<ConstantOp>(op) || isa<DimOp>(op);
}
// Move arguments of the given kernel function into the function if this reduces
// the number of kernel arguments.
static gpu::LaunchFuncOp inlineBeneficiaryOps(gpu::GPUFuncOp kernelFunc,
gpu::LaunchFuncOp launch) {
OpBuilder kernelBuilder(kernelFunc.getBody());
auto &firstBlock = kernelFunc.getBody().front();
SmallVector<ValuePtr, 8> newLaunchArgs;
BlockAndValueMapping map;
for (int i = 0, e = launch.getNumKernelOperands(); i < e; ++i) {
map.map(launch.getKernelOperand(i), kernelFunc.getArgument(i));
}
for (int i = launch.getNumKernelOperands() - 1; i >= 0; --i) {
auto operandOp = launch.getKernelOperand(i)->getDefiningOp();
if (!operandOp || !isInliningBeneficiary(operandOp)) {
newLaunchArgs.push_back(launch.getKernelOperand(i));
continue;
}
// Only inline operations that do not create new arguments.
if (!llvm::all_of(operandOp->getOperands(),
[map](ValuePtr value) { return map.contains(value); })) {
continue;
}
auto clone = kernelBuilder.clone(*operandOp, map);
firstBlock.getArgument(i)->replaceAllUsesWith(clone->getResult(0));
firstBlock.eraseArgument(i);
}
if (newLaunchArgs.size() == launch.getNumKernelOperands())
return launch;
std::reverse(newLaunchArgs.begin(), newLaunchArgs.end());
OpBuilder LaunchBuilder(launch);
SmallVector<Type, 8> newArgumentTypes;
newArgumentTypes.reserve(firstBlock.getNumArguments());
for (auto value : firstBlock.getArguments()) {
newArgumentTypes.push_back(value->getType());
}
kernelFunc.setType(LaunchBuilder.getFunctionType(newArgumentTypes, {}));
auto newLaunch = LaunchBuilder.create<gpu::LaunchFuncOp>(
launch.getLoc(), kernelFunc, launch.getGridSizeOperandValues(),
launch.getBlockSizeOperandValues(), newLaunchArgs);
launch.erase();
return newLaunch;
}
// Outline the `gpu.launch` operation body into a kernel function. Replace
// `gpu.return` operations by `std.return` in the generated function.
static gpu::GPUFuncOp outlineKernelFunc(gpu::LaunchOp launchOp) {
Location loc = launchOp.getLoc();
// Create a builder with no insertion point, insertion will happen separately
// due to symbol table manipulation.
OpBuilder builder(launchOp.getContext());
SmallVector<Type, 4> kernelOperandTypes(launchOp.getKernelOperandTypes());
FunctionType type =
FunctionType::get(kernelOperandTypes, {}, launchOp.getContext());
std::string kernelFuncName =
Twine(launchOp.getParentOfType<FuncOp>().getName(), "_kernel").str();
auto outlinedFunc = builder.create<gpu::GPUFuncOp>(loc, kernelFuncName, type);
outlinedFunc.setAttr(gpu::GPUDialect::getKernelFuncAttrName(),
builder.getUnitAttr());
outlinedFunc.body().takeBody(launchOp.body());
injectGpuIndexOperations(loc, outlinedFunc.body());
return outlinedFunc;
}
// Replace `gpu.launch` operations with an `gpu.launch_func` operation launching
// `kernelFunc`. The kernel func contains the body of the `gpu.launch` with
// constant region arguments inlined.
static void convertToLaunchFuncOp(gpu::LaunchOp &launchOp,
gpu::GPUFuncOp kernelFunc) {
OpBuilder builder(launchOp);
auto launchFuncOp = builder.create<gpu::LaunchFuncOp>(
launchOp.getLoc(), kernelFunc, launchOp.getGridSizeOperandValues(),
launchOp.getBlockSizeOperandValues(), launchOp.getKernelOperandValues());
inlineBeneficiaryOps(kernelFunc, launchFuncOp);
launchOp.erase();
}
namespace {
/// Pass that moves the kernel of each LaunchOp into its separate nested module.
///
/// This pass moves the kernel code of each LaunchOp into a function created
/// inside a nested module. It also creates an external function of the same
/// name in the parent module.
///
/// The kernel modules are intended to be compiled to a cubin blob independently
/// in a separate pass. The external functions can then be annotated with the
/// symbol of the cubin accessor function.
class GpuKernelOutliningPass : public ModulePass<GpuKernelOutliningPass> {
public:
void runOnModule() override {
SymbolTable symbolTable(getModule());
bool modified = false;
for (auto func : getModule().getOps<FuncOp>()) {
// Insert just after the function.
Block::iterator insertPt(func.getOperation()->getNextNode());
func.walk([&](gpu::LaunchOp op) {
gpu::GPUFuncOp outlinedFunc = outlineKernelFunc(op);
// Create nested module and insert outlinedFunc. The module will
// originally get the same name as the function, but may be renamed on
// insertion into the parent module.
auto kernelModule = createKernelModule(outlinedFunc, symbolTable);
symbolTable.insert(kernelModule, insertPt);
// Potentially changes signature, pulling in constants.
convertToLaunchFuncOp(op, outlinedFunc);
modified = true;
});
}
// If any new module was inserted in this module, annotate this module as
// a container module.
if (modified)
getModule().setAttr(gpu::GPUDialect::getContainerModuleAttrName(),
UnitAttr::get(&getContext()));
}
private:
// Returns a module containing kernelFunc and all callees (recursive).
ModuleOp createKernelModule(gpu::GPUFuncOp kernelFunc,
const SymbolTable &parentSymbolTable) {
auto context = getModule().getContext();
Builder builder(context);
auto kernelModule =
ModuleOp::create(builder.getUnknownLoc(), kernelFunc.getName());
kernelModule.setAttr(gpu::GPUDialect::getKernelModuleAttrName(),
builder.getUnitAttr());
SymbolTable symbolTable(kernelModule);
symbolTable.insert(kernelFunc);
SmallVector<Operation *, 8> symbolDefWorklist = {kernelFunc};
while (!symbolDefWorklist.empty()) {
if (Optional<SymbolTable::UseRange> symbolUses =
SymbolTable::getSymbolUses(symbolDefWorklist.pop_back_val())) {
for (SymbolTable::SymbolUse symbolUse : *symbolUses) {
StringRef symbolName =
symbolUse.getSymbolRef().cast<FlatSymbolRefAttr>().getValue();
if (symbolTable.lookup(symbolName))
continue;
Operation *symbolDefClone =
parentSymbolTable.lookup(symbolName)->clone();
symbolDefWorklist.push_back(symbolDefClone);
symbolTable.insert(symbolDefClone);
}
}
}
return kernelModule;
}
};
} // namespace
std::unique_ptr<OpPassBase<ModuleOp>> mlir::createGpuKernelOutliningPass() {
return std::make_unique<GpuKernelOutliningPass>();
}
static PassRegistration<GpuKernelOutliningPass>
pass("gpu-kernel-outlining",
"Outline gpu.launch bodies to kernel functions.");
|
