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
|
"""
A simple script to benchmark the OpenCL low-level routines in comparison to the
CPU ones.
"""
from __future__ import print_function, division
import os
import timeit
import numpy as np
from dtcwt.coeffs import biort, qshift
from dtcwt.opencl.lowlevel import NoCLPresentError, get_default_queue
mandrill = np.load(os.path.join(os.path.dirname(__file__), '..', 'tests', 'mandrill.npz'))['mandrill']
h0o, g0o, h1o, g1o = biort('near_sym_b')
h0a, h0b, g0a, g0b, h1a, h1b, g1a, g1b = qshift('qshift_d')
def format_time(t):
units = (
(60*60, 'hr'), (60, 'min'), (1, 's'), (1e-3, 'ms')
)
for scale, unit in units:
if t >= scale:
return '{0:.2f} {1}'.format(t/scale, unit)
return '{0:.2f} {1}'.format(t*1e6, 'us')
def benchmark(statement='pass', setup='pass'):
number, repeat = (1, 3)
min_time = 0
try:
while min_time < 0.2:
number *= 10
times = timeit.repeat(statement, setup, repeat=repeat, number=number)
min_time = min(times)
except NoCLPresentError:
print('Skipping benchmark since OpenCL is not present')
return 1
t = min_time / number
print('{0} loops, best of {1}: {2}'.format(number, repeat, format_time(t)))
return t
def main():
try:
queue = get_default_queue()
print('Using context: {0}'.format(queue.context))
except NoCLPresentError:
print('Skipping OpenCL benchmark since OpenCL is not present')
print('Running NumPy colfilter...')
a = benchmark('colfilter(mandrill, h1o)',
'from dtcwt.numpy.lowlevel import colfilter; from __main__ import mandrill, h1o')
print('Running OpenCL colfilter...')
b = benchmark('colfilter(mandrill, h1o)',
'from dtcwt.opencl.lowlevel import colfilter; from __main__ import mandrill, h1o')
print('Speed up: x{0:.2f}'.format(a/b))
print('=====')
print('Running NumPy coldfilt...')
a = benchmark('coldfilt(mandrill, h0b, h0a)',
'from dtcwt.numpy.lowlevel import coldfilt; from __main__ import mandrill, h0b, h0a')
print('Running OpenCL coldfilt...')
b = benchmark('coldfilt(mandrill, h0b, h0a)',
'from dtcwt.opencl.lowlevel import coldfilt; from __main__ import mandrill, h0b, h0a')
print('Speed up: x{0:.2f}'.format(a/b))
print('=====')
print('Running NumPy colifilt...')
a = benchmark('colifilt(mandrill, h0b, h0a)',
'from dtcwt.numpy.lowlevel import colifilt; from __main__ import mandrill, h0b, h0a')
print('Running OpenCL colifilt...')
b = benchmark('colifilt(mandrill, h0b, h0a)',
'from dtcwt.opencl.lowlevel import colifilt; from __main__ import mandrill, h0b, h0a')
print('Speed up: x{0:.2f}'.format(a/b))
print('=====')
print('Running NumPy dtwavexfm2...')
a = benchmark('dtwavexfm2(mandrill)',
'from dtcwt.compat import dtwavexfm2; from __main__ import mandrill')
print('Running OpenCL dtwavexfm2...')
b = benchmark('dtwavexfm2(mandrill)',
'from dtcwt.opencl.transform2d import dtwavexfm2; from __main__ import mandrill')
print('Speed up: x{0:.2f}'.format(a/b))
print('=====')
print('Running NumPy dtwavexfm2 (non-POT)...')
a = benchmark('dtwavexfm2(mandrill[:510,:480])',
'from dtcwt.compat import dtwavexfm2; from __main__ import mandrill')
print('Running OpenCL dtwavexfm2 (non-POT)...')
b = benchmark('dtwavexfm2(mandrill[:510,:480])',
'from dtcwt.opencl.transform2d import dtwavexfm2; from __main__ import mandrill')
print('Speed up: x{0:.2f}'.format(a/b))
print('=====')
if __name__ == '__main__':
main()
|
