blob: 0ab90dccc277916869560e4904cce63289d5a3b3 (
plain)
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
|
# This tests the amount of overhead added for inline() function calls.
# It isn't a "real world" test, but is somewhat informative.
# C:\home\ej\wrk\scipy\compiler\examples>python py_none.py
# python: 0.0199999809265
# inline: 0.160000085831
# speed up: 0.124999813736 (this is about a factor of 8 slower)
import time
import sys
sys.path.insert(0,'..')
from inline_tools import inline
def py_func():
return None
n = 10000
t1 = time.time()
for i in range(n):
py_func()
t2 = time.time()
py_time = t2 - t1
print 'python:', py_time
inline("",[])
t1 = time.time()
for i in range(n):
inline("",[])
t2 = time.time()
print 'inline:', (t2-t1)
print 'speed up:', py_time/(t2-t1)
print 'or (more likely) slow down:', (t2-t1)/py_time
|
