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
|
#!/usr/bin/env python
"""Run benchmarks for model synthetic photometry"""
import time
from collections import OrderedDict
import numpy as np
import sncosmo
delim = 61 * "-"
# test data
ndata = 100 # make divisible by 4!
dates = np.linspace(-15., 40., ndata)
bands = np.array((ndata//4) * ['desg', 'desr', 'desi', 'sdssg'])
niter = 100
# models
f99dust = sncosmo.F99Dust(3.1)
models = OrderedDict([
('salt2', sncosmo.Model(source='salt2')),
('hsiao', sncosmo.Model(source='hsiao')),
('salt2+f99dust',
sncosmo.Model(source='salt2', effects=[f99dust],
effect_names=['mw'], effect_frames=['obs'])),
('hsiao+f99dust',
sncosmo.Model(source='hsiao', effects=[f99dust],
effect_names=['mw'], effect_frames=['obs']))
])
print("\nbandflux(band_array, time_array) [4 des bands]:")
print(delim)
print("Model n=1 n=10 n=100")
print(delim)
for name, model in models.items():
print('{:15s}'.format(name), end='')
for idx in [0, range(10), range(100)]:
d = dates[idx]
b = bands[idx]
time1 = time.time()
for i in range(niter): model.bandflux(b, d)
time2 = time.time()
time_sec = (time2 - time1) / niter
print("%10.5f" % (time_sec * 1000.), end='')
print(" ms per call")
|
