[Scipy-tickets] [SciPy] #1159: Wrong eigenvalues generated by scipy.linalg.eigh

SciPy Trac scipy-tickets@scipy....
Thu Apr 22 10:06:30 CDT 2010


#1159: Wrong eigenvalues generated by scipy.linalg.eigh
------------------------------------------------+---------------------------
 Reporter:  ysr                                 |       Owner:  somebody      
     Type:  defect                              |      Status:  needs_decision
 Priority:  normal                              |   Milestone:  0.8.0         
Component:  scipy.linalg                        |     Version:  0.7.0         
 Keywords:  eigenvalue, hermitian, eigenvector  |  
------------------------------------------------+---------------------------

Comment(by ysr):

 Thanks for coding it up in fortran and forwarding it to the Lapack people!

 Yes, I think arpack does not have this problem (since MATLAB works
 correctly on it and they use the Arnoldi Iterations). I will try it in
 Octave. There should be an eigs/eigh built on top of arpack in scipy. Can
 scipy 0.7.2 get this feature? Computing wrong eigenvalues is pretty
 horrific for a linear algebra library.

-- 
Ticket URL: <http://projects.scipy.org/scipy/ticket/1159#comment:4>
SciPy <http://www.scipy.org>
SciPy is open-source software for mathematics, science, and engineering.


More information about the Scipy-tickets mailing list