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
|
"""Compare the speed of exact one-norm calculation vs. its estimation.
"""
from __future__ import division, print_function, absolute_import
import time
import numpy as np
from numpy.testing import (Tester, TestCase, assert_allclose)
import scipy.sparse
class BenchmarkOneNormEst(TestCase):
def bench_onenormest(self):
# print headers and define the column formats
print()
print(' calculation and estimation of one-norm of matrix squaring')
print('==============================================================')
print(' shape | repeats | operation | time ')
print(' | (seconds)')
print('--------------------------------------------------------------')
fmt = ' %15s | %3d | %18s | %6.2f '
np.random.seed(1234)
nrepeats = 100
for n in (2, 3, 5, 10, 30, 100, 300, 500, 1000):
shape = (n, n)
# Sample the matrices.
tm_start = time.clock()
matrices = []
for i in range(nrepeats):
M = np.random.randn(*shape)
matrices.append(M)
tm_end = time.clock()
tm_sampling = tm_end - tm_start
# Get the exact values of one-norms of squares.
tm_start = time.clock()
for M in matrices:
M2 = M.dot(M)
scipy.sparse.linalg.matfuncs._onenorm(M)
tm_end = time.clock()
tm_exact = tm_end - tm_start
# Get the estimates of one-norms of squares.
tm_start = time.clock()
for M in matrices:
scipy.sparse.linalg.matfuncs._onenormest_matrix_power(M, 2)
tm_end = time.clock()
tm_estimate = tm_end - tm_start
# write the rows
print(fmt % (shape, nrepeats, 'matrix sampling', tm_sampling))
print(fmt % (shape, nrepeats, 'one-norm exact', tm_exact))
print(fmt % (shape, nrepeats, 'one-norm estimate', tm_estimate))
print()
if __name__ == '__main__':
Tester().bench()
|
