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
|
#include "benchmark/benchmark.h"
#include <assert.h>
#include <math.h>
#include <stdint.h>
#include <cstdlib>
#include <iostream>
#include <limits>
#include <list>
#include <map>
#include <mutex>
#include <set>
#include <sstream>
#include <string>
#include <vector>
#include <chrono>
#include <thread>
#include <utility>
#if defined(__GNUC__)
# define BENCHMARK_NOINLINE __attribute__((noinline))
#else
# define BENCHMARK_NOINLINE
#endif
namespace {
int BENCHMARK_NOINLINE Factorial(uint32_t n) {
return (n == 1) ? 1 : n * Factorial(n - 1);
}
double CalculatePi(int depth) {
double pi = 0.0;
for (int i = 0; i < depth; ++i) {
double numerator = static_cast<double>(((i % 2) * 2) - 1);
double denominator = static_cast<double>((2 * i) - 1);
pi += numerator / denominator;
}
return (pi - 1.0) * 4;
}
std::set<int> ConstructRandomSet(int size) {
std::set<int> s;
for (int i = 0; i < size; ++i)
s.insert(i);
return s;
}
std::mutex test_vector_mu;
std::vector<int>* test_vector = nullptr;
} // end namespace
static void BM_Factorial(benchmark::State& state) {
int fac_42 = 0;
while (state.KeepRunning())
fac_42 = Factorial(8);
// Prevent compiler optimizations
std::stringstream ss;
ss << fac_42;
state.SetLabel(ss.str());
}
BENCHMARK(BM_Factorial);
BENCHMARK(BM_Factorial)->UseRealTime();
static void BM_CalculatePiRange(benchmark::State& state) {
double pi = 0.0;
while (state.KeepRunning())
pi = CalculatePi(state.range(0));
std::stringstream ss;
ss << pi;
state.SetLabel(ss.str());
}
BENCHMARK_RANGE(BM_CalculatePiRange, 1, 1024 * 1024);
static void BM_CalculatePi(benchmark::State& state) {
static const int depth = 1024;
while (state.KeepRunning()) {
benchmark::DoNotOptimize(CalculatePi(depth));
}
}
BENCHMARK(BM_CalculatePi)->Threads(8);
BENCHMARK(BM_CalculatePi)->ThreadRange(1, 32);
BENCHMARK(BM_CalculatePi)->ThreadPerCpu();
static void BM_SetInsert(benchmark::State& state) {
while (state.KeepRunning()) {
state.PauseTiming();
std::set<int> data = ConstructRandomSet(state.range(0));
state.ResumeTiming();
for (int j = 0; j < state.range(1); ++j)
data.insert(rand());
}
state.SetItemsProcessed(state.iterations() * state.range(1));
state.SetBytesProcessed(state.iterations() * state.range(1) * sizeof(int));
}
BENCHMARK(BM_SetInsert)->Ranges({{1<<10,8<<10}, {1,10}});
template<typename Container, typename ValueType = typename Container::value_type>
static void BM_Sequential(benchmark::State& state) {
ValueType v = 42;
while (state.KeepRunning()) {
Container c;
for (int i = state.range(0); --i; )
c.push_back(v);
}
const size_t items_processed = state.iterations() * state.range(0);
state.SetItemsProcessed(items_processed);
state.SetBytesProcessed(items_processed * sizeof(v));
}
BENCHMARK_TEMPLATE2(BM_Sequential, std::vector<int>, int)->Range(1 << 0, 1 << 10);
BENCHMARK_TEMPLATE(BM_Sequential, std::list<int>)->Range(1 << 0, 1 << 10);
// Test the variadic version of BENCHMARK_TEMPLATE in C++11 and beyond.
#if __cplusplus >= 201103L
BENCHMARK_TEMPLATE(BM_Sequential, std::vector<int>, int)->Arg(512);
#endif
static void BM_StringCompare(benchmark::State& state) {
std::string s1(state.range(0), '-');
std::string s2(state.range(0), '-');
while (state.KeepRunning())
benchmark::DoNotOptimize(s1.compare(s2));
}
BENCHMARK(BM_StringCompare)->Range(1, 1<<20);
static void BM_SetupTeardown(benchmark::State& state) {
if (state.thread_index == 0) {
// No need to lock test_vector_mu here as this is running single-threaded.
test_vector = new std::vector<int>();
}
int i = 0;
while (state.KeepRunning()) {
std::lock_guard<std::mutex> l(test_vector_mu);
if (i%2 == 0)
test_vector->push_back(i);
else
test_vector->pop_back();
++i;
}
if (state.thread_index == 0) {
delete test_vector;
}
}
BENCHMARK(BM_SetupTeardown)->ThreadPerCpu();
static void BM_LongTest(benchmark::State& state) {
double tracker = 0.0;
while (state.KeepRunning()) {
for (int i = 0; i < state.range(0); ++i)
benchmark::DoNotOptimize(tracker += i);
}
}
BENCHMARK(BM_LongTest)->Range(1<<16,1<<28);
static void BM_ParallelMemset(benchmark::State& state) {
int size = state.range(0) / sizeof(int);
int thread_size = size / state.threads;
int from = thread_size * state.thread_index;
int to = from + thread_size;
if (state.thread_index == 0) {
test_vector = new std::vector<int>(size);
}
while (state.KeepRunning()) {
for (int i = from; i < to; i++) {
// No need to lock test_vector_mu as ranges
// do not overlap between threads.
benchmark::DoNotOptimize(test_vector->at(i) = 1);
}
}
if (state.thread_index == 0) {
delete test_vector;
}
}
BENCHMARK(BM_ParallelMemset)->Arg(10 << 20)->ThreadRange(1, 4);
static void BM_ManualTiming(benchmark::State& state) {
size_t slept_for = 0;
int microseconds = state.range(0);
std::chrono::duration<double, std::micro> sleep_duration {
static_cast<double>(microseconds)
};
while (state.KeepRunning()) {
auto start = std::chrono::high_resolution_clock::now();
// Simulate some useful workload with a sleep
std::this_thread::sleep_for(std::chrono::duration_cast<
std::chrono::nanoseconds>(sleep_duration));
auto end = std::chrono::high_resolution_clock::now();
auto elapsed =
std::chrono::duration_cast<std::chrono::duration<double>>(
end - start);
state.SetIterationTime(elapsed.count());
slept_for += microseconds;
}
state.SetItemsProcessed(slept_for);
}
BENCHMARK(BM_ManualTiming)->Range(1, 1 << 14)->UseRealTime();
BENCHMARK(BM_ManualTiming)->Range(1, 1 << 14)->UseManualTime();
#if __cplusplus >= 201103L
template <class ...Args>
void BM_with_args(benchmark::State& state, Args&&...) {
while (state.KeepRunning()) {}
}
BENCHMARK_CAPTURE(BM_with_args, int_test, 42, 43, 44);
BENCHMARK_CAPTURE(BM_with_args, string_and_pair_test,
std::string("abc"), std::pair<int, double>(42, 3.8));
void BM_non_template_args(benchmark::State& state, int, double) {
while(state.KeepRunning()) {}
}
BENCHMARK_CAPTURE(BM_non_template_args, basic_test, 0, 0);
#endif // __cplusplus >= 201103L
BENCHMARK_MAIN()
|
