Opened 10 years ago

Last modified 10 years ago

#11468 closed defect

Memleak in singular.pyx — at Version 4

Reported by: jpflori Owned by: rlm
Priority: critical Milestone: sage-4.7.2
Component: memleak Keywords: memleak, libsingular
Cc: jpflori, burcin Merged in:
Authors: Reviewers:
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description (last modified by jpflori)

Using the following piece of code makes the memory footprint of sage
grow indefinitely:

sage: K = GF(1<<50,'t')
sage: R.<x,y> = PolynomialRing?(K)

sage: a = K.random_element()
sage: while 1:
....: R(a)
....:

The memleak happens when different si2sa_* functions are called.

See http://groups.google.com/group/sage-support/browse_thread/thread/9a8e887df34a8e9a for further discussion.

Change History (4)

comment:1 Changed 10 years ago by jpflori

  • Description modified (diff)
  • Keywords libsingular added; elliptic curves removed
  • Summary changed from Memleak when using elliptic curves to Memleak in multi_polynomial_libsingular.pyx
  • Type changed from PLEASE CHANGE to defect

comment:2 Changed 10 years ago by jpflori

Calling gc.collect() just after the creation prevents the memory problem.

But it does not if called later.

comment:3 Changed 10 years ago by jpflori

  • Status changed from new to needs_review
  • Summary changed from Memleak in multi_polynomial_libsingular.pyx to Memleak in singular.pyx

comment:4 Changed 10 years ago by jpflori

  • Description modified (diff)

I finally found the memleaks in different si2sa_* functions.

Potential fix provided.

Note: See TracTickets for help on using tickets.