On Wednesday, April 30, 2014 7:19:12 AM UTC-7, pete.d...@port.ac.uk wrote:
>
> Hi all,
>
> I'm pretty new to Python, so perhaps I'm doing something wrong, but I've
> encountered what I believe to be a memory leak in Sage's Cone.dual() method.
> Below is some very simplified proof of concept code demonstrating the issue:
>
> #!/usr/bin/env sage
> generators = []
> generators.append( [1,0,0,0] )
> generators.append( [0,1,0,0] )
> generators.append( [0,0,1,0] )
> generators.append( [0,0,0,1] )
> for i in range( 0, 1000 ):
> cone = Cone( generators )
> dual = cone.dual()
> print( 'Current memory usage: ' + str( get_memory_usage() ) )
>
>
> Memory usage steadily increases with each iteration of the for loop.
> Commenting out the line dual = cone.dual(), or moving the line cone =
> Cone( generators ) before the for loop keeps memory usage constant. I'm
> currently using Sage 6.0 on Ubuntu 13.10 32-bit, but have seen the same
> behaviour in Sage 6.1.1 on Ubuntu 64-bit.
>
This code:
generators = [[1,0,0,0],[0,1,0,0],[0,0,1,0],[0,0,0,1]]
import gc
from collections import Counter
cone = Cone( generators )
dual = cone.dual()
gc.collect()
pre={id(p) for p in gc.get_objects()}
for i in range( 0, 2000 ):
cone = Cone( generators )
dual = cone.dual()
gc.collect()
gc.collect()
gc.collect()
T=Counter([str(type(p)) for p in gc.get_objects() if id(p) not in pre])
[(t,c) for t,c in T.iteritems() if c > 10]
confirms a memory leak in, e.g., 6.0 (you see all kinds of objects pile up
in memory), but in 6.2.rc0 there is not, and I see flat memory usage. So it
seems the leak is fixed in a future version.
--
You received this message because you are subscribed to the Google Groups
"sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-support+unsubscr...@googlegroups.com.
To post to this group, send email to sage-support@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-support.
For more options, visit https://groups.google.com/d/optout.
