On Fri, 13 Jul 2007 16:18:38 +0300, Marc 'BlackJack' Rintsch
<[EMAIL PROTECTED]> wrote:
>> $ python -m timeit -c 'import csv; data =
>> list(csv.reader(open("some.csv",
>> "rb")))'
>> 10000 loops, best of 3: 44 usec per loop
>> $ python -m timeit -c 'import csv; data = [row for row in
>> csv.reader(open("some.csv", "rb"))]'
>> 10000 loops, best of 3: 37 usec per loop
>>
>> I don't know why there seems to be a differece, but I know that list
>> comps
>> are python are very heavily optimised.
>
> Does the machine use power saving features like SpeedStep or
> something similar, i.e. runs the processor always with 100% speed or is
> it
> dynamically stepped if there's load on the processor? Do both tests read
> the data always from cache or has the very first loop had to fetch the
> CSV
> file from disk?
>
> $ python -m timeit -n 1000 -c 'import csv; data = [row for row in
> csv.reader(open("test.csv", "rb"))]' 1000 loops, best of 3: 1.27 msec per
> loop
>
> $ python -m timeit -n 1000 -c 'import csv; data =
> list(csv.reader(open("test.csv", "rb")))' 1000 loops, best of 3: 1.25
> msec
> per loop
No SpeedStep - tried a few repeats just in case files were cached,
consistent 35usec for comp 40usec for list
Python 2.5.1 on Linux 1.2ghz
Even replacing the csv lookup with a straight variable declaration:
[range(10)*3], same results
Weird.
Python
--
http://mail.python.org/mailman/listinfo/python-list
