[NumPy-Tickets] [NumPy] #2048: memory overflow when using numpy load in a loop

[NumPy Trac]

#2048: memory overflow when using numpy load in a loop
---------------------------------------------------+------------------------
 Reporter:  eldada                                 |       Owner:  somebody   
     Type:  defect                                 |      Status:  new        
 Priority:  normal                                 |   Milestone:  Unscheduled
Component:  numpy.lib                              |     Version:  1.5.1      
 Keywords:  load, memory, leak, overflow, NpzFile  |  
---------------------------------------------------+------------------------

Comment(by eldada):

 I have posted the question on http://stackoverflow.com/a/9245218/1204331
 and received an answer with a solution + a general way to solve similar
 issues.
  The specific solution is to delete data.f
 After testing the solution proposed by pv (to gc.collect in the loop) I
 found that the del data.f solution is about 2000 times faster. The other
 proposed solution by pv (data.f=None) is as fast as deletion. I wonder why
 would the NpzFile class would be coded in a cyclic referring manner.

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/2048#comment:2>
NumPy <http://projects.scipy.org/numpy>
My example project