List problem

[Josiah Carlson]

bokr at oz.net (Bengt Richter) wrote:

>  delsome_br
>  3.39048024526
>  3.63548729364
>  3.6326486655
>  delsome_am
>  6.40547292869
>  6.32403355062
>  6.21752171923

FYI: list comprehensions in 2.4 are significantly faster than list
comprehensions in 2.3, which is likely where a large portion of your
difference is coming from.

I wasn't sure that I believed your statements in regards to Python's
cache performance, so I thought I'd give it a look...

Python 2.3.2 (#49, Oct  2 2003, 20:02:00) [MSC v.1200 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> a = range(1000000)
>>> import time
>>> def f(a):
...     t = time.time()
...     for i in a:
...             pass
...     return time.time()-t
...
>>> for i in xrange(10):
...     print f(a),
...
0.421999931335 0.421000003815 0.421999931335 0.422000169754 0.421999931335 0.437
99996376 0.421000003815 0.421999931335 0.43799996376 0.406000137329
>>> import random
>>> random.shuffle(a)
>>> for i in xrange(10):
...     print f(a),
...
0.594000101089 0.59299993515 0.610000133514 0.608999967575 0.608999967575 0.5940
00101089 0.593999862671 0.608999967575 0.594000101089 0.593999862671
>>> a.sort()
>>> for i in xrange(10):
...     print f(a),
...
0.421999931335 0.421999931335 0.421999931335 0.422000169754 0.43700003624 0.4219
99931335 0.421999931335 0.421999931335 0.422000169754 0.43700003624


Looks to be an almost 50% slowdown on my dual celeron 400mhz machine
(128kb cache).  Very interesting, I would have never have expected as
much (due to the amount of overhead of Python data structures, I
believed it was generally a wash).

 - Josiah