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
|
#include <stdio.h>
#include <string.h>
#include "catch2/catch_all.hpp"
#include "Recast.h"
#include "RecastAlloc.h"
#include "RecastAssert.h"
#include <vector>
// TODO: Implement benchmarking for platforms other than posix.
#ifdef __unix__
#include <unistd.h>
#ifdef _POSIX_TIMERS
#include <time.h>
#include <stdint.h>
int64_t NowNanos() {
struct timespec tp;
clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &tp);
return tp.tv_nsec + 1000000000LL * tp.tv_sec;
}
#define BM(name, iterations) \
struct BM_ ## name { \
static void Run() { \
int64_t begin_time = NowNanos(); \
for (int i = 0 ; i < iterations; i++) { \
Body(); \
} \
int64_t nanos = NowNanos() - begin_time; \
printf("BM_%-35s %ld iterations in %10ld nanos: %10.2f nanos/it\n", #name ":", (int64_t)iterations, nanos, double(nanos) / iterations); \
} \
static void Body(); \
}; \
TEST_CASE(#name) { \
BM_ ## name::Run(); \
} \
void BM_ ## name::Body()
const int64_t kNumLoops = 100;
const int64_t kNumInserts = 100000;
// Prevent compiler from eliding a calculation.
// TODO: Implement for MSVC.
template <typename T>
void DoNotOptimize(T* v) {
asm volatile ("" : "+r" (v));
}
BM(FlatArray_Push, kNumLoops)
{
int cap = 64;
int* v = (int*)rcAlloc(cap * sizeof(int), RC_ALLOC_TEMP);
for (int j = 0; j < kNumInserts; j++) {
if (j == cap) {
cap *= 2;
int* tmp = (int*)rcAlloc(sizeof(int) * cap, RC_ALLOC_TEMP);
memcpy(tmp, v, j * sizeof(int));
rcFree(v);
v = tmp;
}
v[j] = 2;
}
DoNotOptimize(v);
rcFree(v);
}
BM(FlatArray_Fill, kNumLoops)
{
int* v = (int*)rcAlloc(sizeof(int) * kNumInserts, RC_ALLOC_TEMP);
for (int j = 0; j < kNumInserts; j++) {
v[j] = 2;
}
DoNotOptimize(v);
rcFree(v);
}
BM(FlatArray_Memset, kNumLoops)
{
int* v = (int*)rcAlloc(sizeof(int) * kNumInserts, RC_ALLOC_TEMP);
memset(v, 0, kNumInserts * sizeof(int));
DoNotOptimize(v);
rcFree(v);
}
BM(rcVector_Push, kNumLoops)
{
rcTempVector<int> v;
for (int j = 0; j < kNumInserts; j++) {
v.push_back(2);
}
DoNotOptimize(v.data());
}
BM(rcVector_PushPreallocated, kNumLoops)
{
rcTempVector<int> v;
v.reserve(kNumInserts);
for (int j = 0; j < kNumInserts; j++) {
v.push_back(2);
}
DoNotOptimize(v.data());
}
BM(rcVector_Assign, kNumLoops)
{
rcTempVector<int> v;
v.assign(kNumInserts, 2);
DoNotOptimize(v.data());
}
BM(rcVector_AssignIndices, kNumLoops)
{
rcTempVector<int> v;
v.resize(kNumInserts);
for (int j = 0; j < kNumInserts; j++) {
v[j] = 2;
}
DoNotOptimize(v.data());
}
BM(rcVector_Resize, kNumLoops)
{
rcTempVector<int> v;
v.resize(kNumInserts, 2);
DoNotOptimize(v.data());
}
BM(stdvector_Push, kNumLoops)
{
std::vector<int> v;
for (int j = 0; j < kNumInserts; j++) {
v.push_back(2);
}
DoNotOptimize(v.data());
}
BM(stdvector_PushPreallocated, kNumLoops)
{
std::vector<int> v;
v.reserve(kNumInserts);
for (int j = 0; j < kNumInserts; j++) {
v.push_back(2);
}
DoNotOptimize(v.data());
}
BM(stdvector_Assign, kNumLoops)
{
std::vector<int> v;
v.assign(kNumInserts, 2);
DoNotOptimize(v.data());
}
BM(stdvector_AssignIndices, kNumLoops)
{
std::vector<int> v;
v.resize(kNumInserts);
for (int j = 0; j < kNumInserts; j++) {
v[j] = 2;
}
DoNotOptimize(v.data());
}
BM(stdvector_Resize, kNumLoops)
{
std::vector<int> v;
v.resize(kNumInserts, 2);
DoNotOptimize(v.data());
}
#undef BM
#endif // _POSIX_TIMERS
#endif // __unix__
|
