-
Notifications
You must be signed in to change notification settings - Fork 32
Expand file tree
/
Copy pathscampBench.cpp
More file actions
176 lines (135 loc) · 4.46 KB
/
scampBench.cpp
File metadata and controls
176 lines (135 loc) · 4.46 KB
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
// Copyright (c) 2019 Shapelets.io
//
// This Source Code Form is subject to the terms of the Mozilla Public
// License, v. 2.0. If a copy of the MPL was not distributed with this
// file, You can obtain one at http://mozilla.org/MPL/2.0/.
#include <benchmark/benchmark.h>
#include <khiva/internal/matrixInternal.h>
#include <khiva/matrix.h>
#include <algorithm>
#include <cmath>
#include <iterator>
#include "khivaBenchmark.h"
template <af::Backend BE, int D>
void MatrixProfile(benchmark::State &state) {
af::setBackend(BE);
af::setDevice(D);
auto n = state.range(0);
auto m = state.range(1);
auto tss = af::randu(n, f64);
af::array profile;
af::array index;
af::sync();
while (state.KeepRunning()) {
khiva::matrix::matrixProfile(tss, m, profile, index);
profile.eval();
index.eval();
af::sync();
}
addMemoryCounters(state);
}
template <af::Backend BE, int D>
void MatrixProfile2(benchmark::State &state) {
af::setBackend(BE);
af::setDevice(D);
auto n = state.range(0);
auto m = state.range(1);
auto ta = af::randu(n, f64);
auto tb = af::randu(n, f64);
af::array profile;
af::array index;
af::sync();
while (state.KeepRunning()) {
khiva::matrix::matrixProfile(ta, tb, m, profile, index);
profile.eval();
index.eval();
af::sync();
}
addMemoryCounters(state);
}
template <af::Backend BE, int D>
void MatrixProfileThresh(benchmark::State &state) {
af::setBackend(BE);
af::setDevice(D);
auto n = state.range(0);
auto m = state.range(1);
double threshold = 0.5;
auto tss = af::randu(n, f64);
af::array sumCorrelation;
af::sync();
while (state.KeepRunning()) {
khiva::matrix::matrixProfileThresh(tss, m, threshold, sumCorrelation);
sumCorrelation.eval();
af::sync();
}
addMemoryCounters(state);
}
template <af::Backend BE, int D>
void MatrixProfileThresh2(benchmark::State &state) {
af::setBackend(BE);
af::setDevice(D);
auto n = state.range(0);
auto m = state.range(1);
double threshold = 0.5;
auto ta = af::randu(n, f64);
auto tb = af::randu(n, f64);
af::array sumCorrelation;
af::sync();
while (state.KeepRunning()) {
khiva::matrix::matrixProfileThresh(ta, tb, m, threshold, sumCorrelation);
sumCorrelation.eval();
af::sync();
}
addMemoryCounters(state);
}
template <af::Backend BE, int D>
void MatrixProfileLR(benchmark::State &state) {
af::setBackend(BE);
af::setDevice(D);
auto n = state.range(0);
auto m = state.range(1);
auto tss = af::randu(n, f64);
af::array profileLeft;
af::array profileRight;
af::array indexLeft;
af::array indexRight;
af::sync();
while (state.KeepRunning()) {
khiva::matrix::matrixProfileLR(tss, m, profileLeft, indexLeft, profileRight, indexRight);
profileLeft.eval();
profileRight.eval();
indexLeft.eval();
indexRight.eval();
af::sync();
}
addMemoryCounters(state);
}
void cudaBenchmarks() {
// Empty cuda benchmarks because we want to benchmark only on cpu
}
void openclBenchmarks() {
// Empty opencl benchmarks because we want to benchmark only on cpu
}
void cpuBenchmarks() {
BENCHMARK_TEMPLATE(MatrixProfile, af::Backend::AF_BACKEND_CPU, CPU_BENCHMARKING_DEVICE)
->RangeMultiplier(2)
->Ranges({{16 << 10, 128 << 10}, {16, 512}})
->Unit(benchmark::TimeUnit::kMicrosecond);
BENCHMARK_TEMPLATE(MatrixProfile2, af::Backend::AF_BACKEND_CPU, CPU_BENCHMARKING_DEVICE)
->RangeMultiplier(2)
->Ranges({{16 << 10, 128 << 10}, {16, 512}})
->Unit(benchmark::TimeUnit::kMicrosecond);
BENCHMARK_TEMPLATE(MatrixProfileLR, af::Backend::AF_BACKEND_CPU, CPU_BENCHMARKING_DEVICE)
->RangeMultiplier(2)
->Ranges({{16 << 10, 128 << 10}, {16, 512}})
->Unit(benchmark::TimeUnit::kMicrosecond);
BENCHMARK_TEMPLATE(MatrixProfileThresh, af::Backend::AF_BACKEND_CPU, CPU_BENCHMARKING_DEVICE)
->RangeMultiplier(2)
->Ranges({{16 << 10, 128 << 10}, {16, 512}})
->Unit(benchmark::TimeUnit::kMicrosecond);
BENCHMARK_TEMPLATE(MatrixProfileThresh2, af::Backend::AF_BACKEND_CPU, CPU_BENCHMARKING_DEVICE)
->RangeMultiplier(2)
->Ranges({{16 << 10, 128 << 10}, {16, 512}})
->Unit(benchmark::TimeUnit::kMicrosecond);
}
KHIVA_BENCHMARK_MAIN(cudaBenchmarks, openclBenchmarks, cpuBenchmarks)