[SciPy-User] Eigenvectors of sparse symmetric matrix

Pauli Virtanen pav@iki...
Tue Oct 26 04:05:19 CDT 2010


Mon, 25 Oct 2010 18:31:08 -0700, Lutz Maibaum wrote:

> On Oct 25, 2010, at 5:39 PM, Lutz Maibaum wrote:
>> On Oct 25, 2010, at 3:27 PM, Pauli Virtanen wrote:
>>> The wrappers are currently written to print a warning if the iteration
>>> does not converge, and I do not see yet why that doesn't happen.
>> 
>> Interesting, I don't get a warning. I see that there is a convergence
>> check in _SymmetricArpackParams.iterate(), but not in
>> _UnsymmetricArpackParams.iterate(). Could that be the reason?
> 
> The test is actually there, the symmetric case just has an additional
> test.
> 
> However, I noticed that the warning is triggered when the returned
> "info" status is -1. Is it possible that it should be compared to 1?
> From http://www.caam.rice.edu/software/ARPACK/UG/node137.html:
[clip]
> c  INFO    Integer.  (INPUT/OUTPUT)
> c          If INFO .EQ. 0, a randomly initial residual vector is used. 
> c          If INFO .NE. 0, RESID contains the initial residual vector, 
> c                          possibly from a previous run. 
> c          Error flag on output.
> c          =  0: Normal exit.
> c          =  1: Maximum number of iterations taken. 
[clip]

Seems indeed that the check is wrong. Needs fixing.

-- 
Pauli Virtanen



More information about the SciPy-User mailing list