[Pyrex] Possible Mem Leak...

Alexander A Naanou alex_nanou at pochtamt.ru
Thu Jun 5 01:00:59 CEST 2003


Hi!

  It appears that the following Pyrex code produces a substantial
memory leak.

---cut---

def f(a, *p, **n):
	pass


---uncut---

as can be seen from the graph "pyrex_leak.png" the pyrex version 
of the function (see: test.pyx + pyrex_leak.py) produces a memory
leek, and comparing to the python version (pyrex_leak_ref.py) is
somewhat slower (considering the mem usage, this is not unexpected).

"pyrex_leak_ref.png" is the pure python reference graph.


see attachment for all the data mentioned...

Test System:
Platform: 2xPIII 560MHz, 1024MB RAM (PC133), WinXP
Python: 2.3b1
Pyrex: 0.7.2
C Compiler: GCC 3.2 on Cygwin

NOTE: the GCC compiler was used for all components used in 
      the test... (except the OS :) )


                                     With Respect..
                                               Alex.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_data.tgz
Type: application/x-gzip-compressed
Size: 51281 bytes
Desc: not available
Url : http://lists.copyleft.no/pipermail/pyrex/attachments/20030605/652834f6/test_data.bin


More information about the Pyrex mailing list