[Scipy-tickets] [SciPy] #1717: QZ segfault on Windows for 0.11.0rc1

SciPy Trac scipy-tickets@scipy....
Sat Sep 8 10:41:41 CDT 2012


#1717: QZ segfault on Windows for 0.11.0rc1
---------------------------+------------------------------------------------
  Reporter:  rgommers      |       Owner:  somebody
      Type:  defect        |      Status:  reopened
  Priority:  highest       |   Milestone:  0.11.0  
 Component:  scipy.linalg  |     Version:  devel   
Resolution:                |    Keywords:          
---------------------------+------------------------------------------------
Changes (by rgommers):

  * status:  closed => reopened
  * resolution:  fixed =>


Comment:

 I've done some more extensive testing for this. Some observations:

   - With sort=None, everything is fine. I tried testing with 1000s of
 matrices of random numbers, with sizes from 2x2 to >1e3x1e3.
   - With other sort functions segfaults occur as described above. Also for
 the 4 sort functions that can be specified by a string.
   - I double checked again the LWORK size, that seems correct. I hardcoded
 larger sizes; independent of the size of LWORK segfaults occur.

 The failures reported for rc2 on other platforms were due to a simple
 unrelated issue, see PR-292. So it's still Windows only with ATLAS
 binaries from the numpy vendor branch and SSE2/SSE3.

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1717#comment:22>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list