On Sat, Apr 23, 2011 at 4:17 PM, Christoph Gohlke <cgohlke@uci.edu> wrote:


On 4/23/2011 2:47 PM, Charles R Harris wrote:


On Sat, Apr 23, 2011 at 3:38 PM, Christoph Gohlke <cgohlke@uci.edu
<mailto:cgohlke@uci.edu>> wrote:



   On 4/23/2011 10:41 AM, Charles R Harris wrote:
   >
   >
   >  On Sat, Apr 23, 2011 at 11:09 AM, Bruce Southey
   <bsouthey@gmail.com <mailto:bsouthey@gmail.com>
   >  <mailto:bsouthey@gmail.com <mailto:bsouthey@gmail.com>>> wrote:
   >
   >      On Sat, Apr 23, 2011 at 9:58 AM, Till Stensitzki
   <mail.till@gmx.de <mailto:mail.till@gmx.de>
   >  <mailto:mail.till@gmx.de <mailto:mail.till@gmx.de>>> wrote:
   >  >
   >  > > Do you also have an earlier version of numpy installed? As David
   >      says, this
   >  > >should raise an error for recent numpy and
   >  > >I'm wondering if you are inadvertently
   >  > >running an earlier version.Chuck
   >  >
   >  >
   >  >  I only have one python installation and
   >  >  numpy.__version__ shows 1.6b.
   >  >  I could reinstall numpy, if it would help.
   >  >
   >  >  _______________________________________________
   >  >  NumPy-Discussion mailing list
   >  > NumPy-Discussion@scipy.org <mailto:NumPy-Discussion@scipy.org>
   <mailto:NumPy-Discussion@scipy.org <mailto:NumPy-Discussion@scipy.org>>

   >  > http://mail.scipy.org/mailman/listinfo/numpy-discussion
   >  >
   >
   >      Hi,
   >      I can get this with 64-bit Win 7, 32-bit Python 2.6, 2.7
   (below) and
   >      3.1 and numpy 1.6b (fresh install)  IDLE and the command line.
   I can
   >      also confirm the 'ValueError' with Python2.6 and numpy 1.51 on the
   >      same system.
   >
   >      Actually this is 'weird' when printing and crashed with the
   range -
   >      accessing unassigned memory?
   >      A smaller array gives an numpy error or memory error in idle.
   >
   >      Bruce
   >
   >
   >  > >> import numpy as np
   >  > >> x=np.zeros((262144, 262144))
   >  > >> x
   >      array([], shape=(262144, 262144), dtype=float64)
   >  > >> x[0,0]
   >      2.1453735050108555e-314
   >  > >> x[1:10,1:10]
   >
   >  > >> ================================ RESTART
   >      ================================
   >  > >> import numpy as np
   >
   >  > >> x=np.zeros((26214, 26214))
   >
   >      Traceback (most recent call last):
   >        File "<pyshell#7>", line 1, in <module>
   >          x=np.zeros((26214, 26214))
   >      ValueError: array is too big.
   >  > >>
   >  > >> x=np.zeros((262144, 26214))
   >
   >      Traceback (most recent call last):
   >        File "<pyshell#8>", line 1, in <module>
   >          x=np.zeros((262144, 26214))
   >      MemoryError
   >      _____
   >
   >
   >  This was fixed before, maybe it got broken again. Since this looks
   >  windows specific, I'm guessing it has something to do with the size of
   >  long being 32 bits.
   >
   >  The previous problem was integer overflow when multiplying the
   >  dimensions together to get the array size when repeated divisions
   of the
   >  maximum size should have be used instead.
   >
   >  Chuck

   Could be related to this change:

   <https://github.com/numpy/numpy/commit/fcc6cc73ddcb1fc85446ba9256ac24ecdda6c6d8#L1L1121>


My, that does look suspicious ;) Could you revert that loop and test it out?

There was also a function for doing that check, I don't recall which,
and it should probably be checked to make sure it remains as was.

Chuck




Reverting the change worked. A patch is attached.

import numpy as np
np.__version__
'1.6.0b3'

x=np.zeros((262144, 262144))
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>

ValueError: array is too big.


Thanks, applied.

Chuck