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
|
//===-- BypassSlowDivision.cpp - Bypass slow division ---------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file contains an optimization for div and rem on architectures that
// execute short instructions significantly faster than longer instructions.
// For example, on Intel Atom 32-bit divides are slow enough that during
// runtime it is profitable to check the value of the operands, and if they are
// positive and less than 256 use an unsigned 8-bit divide.
//
//===----------------------------------------------------------------------===//
#define DEBUG_TYPE "bypass-slow-division"
#include "llvm/Transforms/Utils/BypassSlowDivision.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/Function.h"
#include "llvm/IRBuilder.h"
#include "llvm/Instructions.h"
using namespace llvm;
namespace {
struct DivOpInfo {
bool SignedOp;
Value *Dividend;
Value *Divisor;
DivOpInfo(bool InSignedOp, Value *InDividend, Value *InDivisor)
: SignedOp(InSignedOp), Dividend(InDividend), Divisor(InDivisor) {}
};
struct DivPhiNodes {
PHINode *Quotient;
PHINode *Remainder;
DivPhiNodes(PHINode *InQuotient, PHINode *InRemainder)
: Quotient(InQuotient), Remainder(InRemainder) {}
};
}
namespace llvm {
template<>
struct DenseMapInfo<DivOpInfo> {
static bool isEqual(const DivOpInfo &Val1, const DivOpInfo &Val2) {
return Val1.SignedOp == Val2.SignedOp &&
Val1.Dividend == Val2.Dividend &&
Val1.Divisor == Val2.Divisor;
}
static DivOpInfo getEmptyKey() {
return DivOpInfo(false, 0, 0);
}
static DivOpInfo getTombstoneKey() {
return DivOpInfo(true, 0, 0);
}
static unsigned getHashValue(const DivOpInfo &Val) {
return (unsigned)(reinterpret_cast<uintptr_t>(Val.Dividend) ^
reinterpret_cast<uintptr_t>(Val.Divisor)) ^
(unsigned)Val.SignedOp;
}
};
typedef DenseMap<DivOpInfo, DivPhiNodes> DivCacheTy;
}
// insertFastDiv - Substitutes the div/rem instruction with code that checks the
// value of the operands and uses a shorter-faster div/rem instruction when
// possible and the longer-slower div/rem instruction otherwise.
static bool insertFastDiv(Function &F,
Function::iterator &I,
BasicBlock::iterator &J,
IntegerType *BypassType,
bool UseDivOp,
bool UseSignedOp,
DivCacheTy &PerBBDivCache) {
// Get instruction operands
Instruction *Instr = J;
Value *Dividend = Instr->getOperand(0);
Value *Divisor = Instr->getOperand(1);
if (isa<ConstantInt>(Divisor) ||
(isa<ConstantInt>(Dividend) && isa<ConstantInt>(Divisor))) {
// Operations with immediate values should have
// been solved and replaced during compile time.
return false;
}
// Basic Block is split before divide
BasicBlock *MainBB = I;
BasicBlock *SuccessorBB = I->splitBasicBlock(J);
++I; //advance iterator I to successorBB
// Add new basic block for slow divide operation
BasicBlock *SlowBB = BasicBlock::Create(F.getContext(), "",
MainBB->getParent(), SuccessorBB);
SlowBB->moveBefore(SuccessorBB);
IRBuilder<> SlowBuilder(SlowBB, SlowBB->begin());
Value *SlowQuotientV;
Value *SlowRemainderV;
if (UseSignedOp) {
SlowQuotientV = SlowBuilder.CreateSDiv(Dividend, Divisor);
SlowRemainderV = SlowBuilder.CreateSRem(Dividend, Divisor);
} else {
SlowQuotientV = SlowBuilder.CreateUDiv(Dividend, Divisor);
SlowRemainderV = SlowBuilder.CreateURem(Dividend, Divisor);
}
SlowBuilder.CreateBr(SuccessorBB);
// Add new basic block for fast divide operation
BasicBlock *FastBB = BasicBlock::Create(F.getContext(), "",
MainBB->getParent(), SuccessorBB);
FastBB->moveBefore(SlowBB);
IRBuilder<> FastBuilder(FastBB, FastBB->begin());
Value *ShortDivisorV = FastBuilder.CreateCast(Instruction::Trunc, Divisor,
BypassType);
Value *ShortDividendV = FastBuilder.CreateCast(Instruction::Trunc, Dividend,
BypassType);
// udiv/urem because optimization only handles positive numbers
Value *ShortQuotientV = FastBuilder.CreateExactUDiv(ShortDividendV,
ShortDivisorV);
Value *ShortRemainderV = FastBuilder.CreateURem(ShortDividendV,
ShortDivisorV);
Value *FastQuotientV = FastBuilder.CreateCast(Instruction::ZExt,
ShortQuotientV,
Dividend->getType());
Value *FastRemainderV = FastBuilder.CreateCast(Instruction::ZExt,
ShortRemainderV,
Dividend->getType());
FastBuilder.CreateBr(SuccessorBB);
// Phi nodes for result of div and rem
IRBuilder<> SuccessorBuilder(SuccessorBB, SuccessorBB->begin());
PHINode *QuoPhi = SuccessorBuilder.CreatePHI(Instr->getType(), 2);
QuoPhi->addIncoming(SlowQuotientV, SlowBB);
QuoPhi->addIncoming(FastQuotientV, FastBB);
PHINode *RemPhi = SuccessorBuilder.CreatePHI(Instr->getType(), 2);
RemPhi->addIncoming(SlowRemainderV, SlowBB);
RemPhi->addIncoming(FastRemainderV, FastBB);
// Replace Instr with appropriate phi node
if (UseDivOp)
Instr->replaceAllUsesWith(QuoPhi);
else
Instr->replaceAllUsesWith(RemPhi);
Instr->eraseFromParent();
// Combine operands into a single value with OR for value testing below
MainBB->getInstList().back().eraseFromParent();
IRBuilder<> MainBuilder(MainBB, MainBB->end());
Value *OrV = MainBuilder.CreateOr(Dividend, Divisor);
// BitMask is inverted to check if the operands are
// larger than the bypass type
uint64_t BitMask = ~BypassType->getBitMask();
Value *AndV = MainBuilder.CreateAnd(OrV, BitMask);
// Compare operand values and branch
Value *ZeroV = MainBuilder.getInt32(0);
Value *CmpV = MainBuilder.CreateICmpEQ(AndV, ZeroV);
MainBuilder.CreateCondBr(CmpV, FastBB, SlowBB);
// point iterator J at first instruction of successorBB
J = I->begin();
// Cache phi nodes to be used later in place of other instances
// of div or rem with the same sign, dividend, and divisor
DivOpInfo Key(UseSignedOp, Dividend, Divisor);
DivPhiNodes Value(QuoPhi, RemPhi);
PerBBDivCache.insert(std::pair<DivOpInfo, DivPhiNodes>(Key, Value));
return true;
}
// reuseOrInsertFastDiv - Reuses previously computed dividend or remainder if
// operands and operation are identical. Otherwise call insertFastDiv to perform
// the optimization and cache the resulting dividend and remainder.
static bool reuseOrInsertFastDiv(Function &F,
Function::iterator &I,
BasicBlock::iterator &J,
IntegerType *BypassType,
bool UseDivOp,
bool UseSignedOp,
DivCacheTy &PerBBDivCache) {
// Get instruction operands
Instruction *Instr = J;
DivOpInfo Key(UseSignedOp, Instr->getOperand(0), Instr->getOperand(1));
DivCacheTy::iterator CacheI = PerBBDivCache.find(Key);
if (CacheI == PerBBDivCache.end()) {
// If previous instance does not exist, insert fast div
return insertFastDiv(F, I, J, BypassType, UseDivOp, UseSignedOp,
PerBBDivCache);
}
// Replace operation value with previously generated phi node
DivPhiNodes &Value = CacheI->second;
if (UseDivOp) {
// Replace all uses of div instruction with quotient phi node
J->replaceAllUsesWith(Value.Quotient);
} else {
// Replace all uses of rem instruction with remainder phi node
J->replaceAllUsesWith(Value.Remainder);
}
// Advance to next operation
++J;
// Remove redundant operation
Instr->eraseFromParent();
return true;
}
// bypassSlowDivision - This optimization identifies DIV instructions that can
// be profitably bypassed and carried out with a shorter, faster divide.
bool llvm::bypassSlowDivision(Function &F,
Function::iterator &I,
const DenseMap<unsigned int, unsigned int> &BypassWidths) {
DivCacheTy DivCache;
bool MadeChange = false;
for (BasicBlock::iterator J = I->begin(); J != I->end(); J++) {
// Get instruction details
unsigned Opcode = J->getOpcode();
bool UseDivOp = Opcode == Instruction::SDiv || Opcode == Instruction::UDiv;
bool UseRemOp = Opcode == Instruction::SRem || Opcode == Instruction::URem;
bool UseSignedOp = Opcode == Instruction::SDiv ||
Opcode == Instruction::SRem;
// Only optimize div or rem ops
if (!UseDivOp && !UseRemOp)
continue;
// Skip division on vector types, only optimize integer instructions
if (!J->getType()->isIntegerTy())
continue;
// Get bitwidth of div/rem instruction
IntegerType *T = cast<IntegerType>(J->getType());
int bitwidth = T->getBitWidth();
// Continue if bitwidth is not bypassed
DenseMap<unsigned int, unsigned int>::const_iterator BI = BypassWidths.find(bitwidth);
if (BI == BypassWidths.end())
continue;
// Get type for div/rem instruction with bypass bitwidth
IntegerType *BT = IntegerType::get(J->getContext(), BI->second);
MadeChange |= reuseOrInsertFastDiv(F, I, J, BT, UseDivOp,
UseSignedOp, DivCache);
}
return MadeChange;
}
|