[issue4074] Building a list of tuples has non-linear performance

Antoine Pitrou report at bugs.python.org
Sun Dec 14 17:11:54 CET 2008


Antoine Pitrou <pitrou at free.fr> added the comment:

Here is a simple patch implementing Martin's proposal with a few basic
tweaks.

Using Greg's script, we get:

-> without patch:

1000000 2.64134001732
2000000 3.60712885857
3000000 5.40855813026
4000000 6.46308898926
5000000 8.65147781372
6000000 10.3949871063
7000000 12.1376650333
8000000 12.7046239376
9000000 15.4652290344

-> with patch:

1000000 2.692289114
2000000 1.91455483437
3000000 2.02900099754
4000000 1.72369813919
5000000 1.87697696686
6000000 2.08952093124
7000000 1.08168196678
8000000 2.32068109512
9000000 1.1202070713

-> with GC disabled:

1000000 1.6810901165
2000000 0.955595970154
3000000 0.959649085999
4000000 0.933673858643
5000000 0.954123973846
6000000 0.929254055023
7000000 0.901160001755
8000000 0.921751022339
9000000 0.894830942154

----------
keywords: +patch
stage:  -> patch review
Added file: http://bugs.python.org/file12353/gctrigger.patch

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue4074>
_______________________________________


More information about the Python-bugs-list mailing list