LinearNDInterpolator memory consumption

[notmatthancock]

I am executing something like the following:

from numpy import *
from scipy import interpolate as irp

w = 61

t = linspace(-1,1,w)
xx,yy = meshgrid(t,t)
xx,yy = xx.flatten(), yy.flatten() # points for evaluation
P = c_[sign(xx)*abs(xx)**0.65, sign(yy)*abs(yy)**0.65] # points for interpolation
del t

x = random.randint( 0, 255, (2*w**2,w**2) ).astype(uint8) # fake interpolation values - each row corresponds to one interpolation
xo = zeros(x.shape,dtype=uint8) # for storing results
for i in range(x.shape[0]): # memory accumulates here I think
    f = irp.LinearNDInterpolator( P, x[i,:] )
    out = f( xx, yy )
    xo[i,:] = out
    del f, out

on a intel core 2 duo macbook OS X 10.6 with 2GB RAM. The memory requirements to perform this interpolation a single time are not large (~100MB), yet when I perform this loop it fails. Python slowly acquires all my memory as the loop progresses. It doesn't make sense that an operation that only needs something on the order of MB grabs more than a GB and then errors because it can't have more. I receive this error message:

Python(10783) malloc: *** mmap(size=16777216) failed (error code=12) *** error: can't allocate region *** set a breakpoint in malloc_error_break to debug Traceback (most recent call last): File "make_lens.py", line 16, in <module> f = irp.LinearNDInterpolator( P, x[i,:] ) File "interpnd.pyx", line 204, in scipy.interpolate.interpnd.LinearNDInterpolator.__init__ (scipy/interpolate/interpnd.c:3794) File "qhull.pyx", line 1703, in scipy.spatial.qhull.Delaunay.__init__ (scipy/spatial/qhull.c:13267) File "qhull.pyx", line 1432, in scipy.spatial.qhull._QhullUser.__init__ (scipy/spatial/qhull.c:11989) File "qhull.pyx", line 1712, in scipy.spatial.qhull.Delaunay._update (scipy/spatial/qhull.c:13470) File "qhull.pyx", line 526, in scipy.spatial.qhull._Qhull.get_simplex_facet_array (scipy/spatial/qhull.c:5453) File "qhull.pyx", line 594, in scipy.spatial.qhull._Qhull._get_simplex_facet_array (scipy/spatial/qhull.c:6010) MemoryError

I tried this with CloughTocher2DInterpolator as well and received a similar error:

Python(10974) malloc: *** mmap(size=16777216) failed (error code=12)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
QH6080 qhull error (qh_memalloc): insufficient memory to allocate short memory buffer (65536 bytes)
While executing:  | qhull d Qbb Qz Qc Qt
Options selected for Qhull 2012.1 2012/02/18:
run-id 1859014481  delaunay  Qbbound-last  Qz-infinity-point  Qcoplanar-keep
Qtriangulate  _pre-merge  _zero-centrum  Qinterior-keep  Pgood  _max-width  2
Error-roundoff 2.8e-15  _one-merge 1.9e-14  Visible-distance 5.5e-15
U-coplanar-distance 5.5e-15  Width-outside 1.1e-14  _wide-facet 3.3e-14
Last point added to hull was p3457.  Last merge was #4395.
Qhull has finished constructing the hull.
At error exit:
Delaunay triangulation by the convex hull of 3722 points in 3-d:

  Number of input sites and at-infinity: 3722
  Number of Delaunay regions: 0
  Number of non-simplicial Delaunay regions: 3604

Statistics for:  | qhull d Qbb Qz Qc Qt

  Number of points processed: 3722
  Number of hyperplanes created: 14131
  Number of facets in hull: 13798
  Number of distance tests for qhull: 128936
  Number of distance tests for merging: 134142
  Number of distance tests for checking: 60878
  Number of merged facets: 4395
  CPU seconds to compute hull (after input): 0.04895

precision problems (corrected unless 'Q0' or an error)
   4395 coplanar horizon facets for new vertices
Segmentation fault

Why is memory not being released? Is qhull leaky?

[pv]

Please indicate which Scipy version you have installed.

[notmatthancock]

scipy version: 0.13.0
numpy version: 1.7.1

[pv]

This is an issue fixed in 0.13.2.