script Re: comparative performance benchmark Python 1.5.2 - 2.0 - 2.1
Frederic Giacometti
fredg at sdsc.edu
Thu Oct 25 14:59:15 EDT 2001
I submitted the scripts to as sourceforge bug 474992...
http://sourceforge.net/tracker/index.php?func=detail&aid=474992&group_id=5470&atid=105470
The scrips can also be obtained through cvs at:
cvs -d:pserver:anonymous at cvs.JPE.sourceforge.net:/cvsroot/jpe login
cvs -d:pserver:anonymous at cvs.JPE.sourceforge.net:/cvsroot/jpe checkout
perftestpython
Here are the results:
version bench1 nogc bench2
1.5.2 0 %
2.0 48 % 33 % 21 %
2.1.1 53 % 35 % 25 %
2.2 51 % 35 % 23 %
where:
bench1 = original python 1.5.2 script
nogc = bench1 when disabling gc
bench2 = bench1 script 'rewritten' for python 2.0 and latter (string
methods, sre -> pre...)
These numbers represent the increase in user CPU time (see bench.out for
more details).
The results are high, under all angles.
All python versions were build using the standard/default configure
parameters.
The runs are on Tru64, but we get similar results on Solaris.
Can these results be confirmed ? (i.e.: if not, what did we get
wrong...?).
If yes ... could the features responsible for the performance hogs (gc +
???) be set as optional, and taken out of the default build ?
Frederic Giacometti
More information about the Python-list
mailing list