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
|
// -*-C++-*-
#define VML_NODEBUG
#include "vecmathlib.h"
#include <cmath>
#include <cstdlib>
#include <iomanip>
#include <iostream>
#include <string>
#include <sys/time.h>
using namespace std;
using namespace vecmathlib;
#ifndef __has_builtin
# define __has_builtin(x) 0 // Compatibility with non-clang compilers
#endif
typedef unsigned long long ticks;
inline ticks getticks()
{
#if __has_builtin(__builtin_readcyclecounter)
return __builtin_readcyclecounter();
#else
# if defined __x86_64__
ticks a, d;
asm volatile("rdtsc" : "=a" (a), "=d" (d));
return a | (d << 32);
# elif defined __powerpc__
unsigned int tbl, tbu, tbu1;
do {
asm volatile("mftbu %0": "=r"(tbu));
asm volatile("mftb %0": "=r"(tbl));
asm volatile("mftbu %0": "=r"(tbu1));
} while (tbu != tbu1);
return ((unsigned long long)tbu << 32) | tbl;
# endif
#endif
}
inline double elapsed(ticks t1, ticks t0)
{
return t1-t0;
}
double get_sys_time()
{
timeval tp;
gettimeofday(&tp, NULL);
return tp.tv_sec + 1.0e-6 * tp.tv_usec;
}
double measure_tick()
{
ticks const rstart = getticks();
double const wstart = get_sys_time();
while (get_sys_time() - wstart < 0.1) {
// do nothing, just wait
}
ticks const rend = getticks();
double const wend = get_sys_time();
assert(wend-wstart >= 0.09);
return (wend - wstart) / elapsed(rend, rstart);
}
double global_result = 0.0;
template<typename realvec_t>
void save_result(realvec_t result)
{
for (int i=0; i<realvec_t::size; ++i) {
global_result += result[i];
}
}
template<typename T> inline T nop(T x) { return x; }
#define DECLARE_FUNCTOR(func) \
template<typename T> \
struct functor_##func { \
static char const* name() { return #func; } \
T operator()(T x) { return func(x); } \
}
DECLARE_FUNCTOR(nop);
DECLARE_FUNCTOR(sqrt);
DECLARE_FUNCTOR(exp);
DECLARE_FUNCTOR(log);
DECLARE_FUNCTOR(sin);
DECLARE_FUNCTOR(cos);
DECLARE_FUNCTOR(atan);
template<typename realvec_t, template<typename> class func_t>
double run_bench()
{
realvec_t x0, dx;
for (int i=0; i<realvec_t::size; ++i) {
x0.set_elt(i, 1.0f + float(i));
dx.set_elt(i, 1.0e-6f);
}
realvec_t x, y;
ticks t0, t1;
double const cycles_per_tick = 1.0; // measure_tick();
int const numiters = 10000000;
func_t<realvec_t> func;
t0 = getticks();
x = y = x0;
for (int n=0; n<numiters; ++n) {
y += func(x);
x += dx;
}
t1 = getticks();
save_result(y);
return cycles_per_tick * elapsed(t1,t0) * realvec_t::size / numiters;
}
template<typename realvec_t, template<typename> class func_t>
void bench_type_func()
{
cout << " "
<< setw(-5) << func_t<realvec_t>::name() << " "
<< setw(18) << realvec_t::name() << ": " << flush;
double const cycles = run_bench<realvec_t, func_t>();
cout << cycles << " cycles\n" << flush;
}
template<template<typename> class func_t>
void bench_func()
{
cout << "\n"
<< "Benchmarking " << func_t<float>().name() << ":\n";
bench_type_func<realpseudovec<float,1>, func_t>();
// bench_type_func<realbuiltinvec<float,1>, func_t>();
bench_type_func<realtestvec<float,1>, func_t>();
#ifdef VECMATHLIB_HAVE_VEC_FLOAT_1
bench_type_func<realvec<float,1>, func_t>();
#endif
#ifdef VECMATHLIB_HAVE_VEC_FLOAT_4
bench_type_func<realpseudovec<float,4>, func_t>();
// bench_type_func<realbuiltinvec<float,4>, func_t>();
bench_type_func<realtestvec<float,4>, func_t>();
bench_type_func<realvec<float,4>, func_t>();
#endif
#ifdef VECMATHLIB_HAVE_VEC_FLOAT_8
bench_type_func<realpseudovec<float,8>, func_t>();
// bench_type_func<realbuiltinvec<float,8>, func_t>();
bench_type_func<realtestvec<float,8>, func_t>();
bench_type_func<realvec<float,8>, func_t>();
#endif
bench_type_func<realpseudovec<double,1>, func_t>();
// bench_type_func<realbuiltinvec<double,1>, func_t>();
bench_type_func<realtestvec<double,1>, func_t>();
#ifdef VECMATHLIB_HAVE_VEC_DOUBLE_1
bench_type_func<realvec<double,1>, func_t>();
#endif
#ifdef VECMATHLIB_HAVE_VEC_DOUBLE_2
bench_type_func<realpseudovec<double,2>, func_t>();
// bench_type_func<realbuiltinvec<double,2>, func_t>();
bench_type_func<realtestvec<double,2>, func_t>();
bench_type_func<realvec<double,2>, func_t>();
#endif
#ifdef VECMATHLIB_HAVE_VEC_DOUBLE_4
bench_type_func<realpseudovec<double,4>, func_t>();
// bench_type_func<realbuiltinvec<double,4>, func_t>();
bench_type_func<realtestvec<double,4>, func_t>();
bench_type_func<realvec<double,4>, func_t>();
#endif
}
void bench()
{
bench_func<functor_nop>();
bench_func<functor_sqrt>();
bench_func<functor_exp>();
bench_func<functor_log>();
bench_func<functor_sin>();
bench_func<functor_cos>();
bench_func<functor_atan>();
}
int main(int argc, char** argv)
{
using namespace vecmathlib;
cout << "Benchmarking math functions:\n";
bench();
// Checking global accumulator to prevent optimisation
if (! std::isfinite(global_result)) {
cout << "\n"
<< "WARNING: Global accumulator is not finite\n";
}
return 0;
}
|
