[SciPy-user] singular value decomposition

Chris Lee c.j.lee@tnw.utwente...
Wed Sep 17 04:33:39 CDT 2008


OK, it turns out that if I use a randomly generated matrix I get about  
a 1 minute svd, which corresponds nicely to the value Robert found.

After a bit of sleuthing, I discovered that the zero packing to make  
the matrix have dimensions that were a power of two were the culprit.  
With the packing removed everything suddenly started working again.  
Now I just have to remove all the important bugs (like why I get  
nonphysical results and noncovergence) :)
Thanks for all your help

Cheers
Chris


On Sep 17, 2008, at 10:12 AM, David Cournapeau wrote:

> Chris Lee wrote:
>> Hi All,
>>
>> I have code that needs to repeatedly perform an svd on a matrix  
>> (2304,
>> 2304) shape. It seems to take forever to do (I haven't timed it, but
>> approximately two hours passed on the first trial (without  
>> completion)
>> and I am now running a second to make sure nothing went wrong.
>>
>> Is this because I am not linked to an optimized BLAS/LAPACK library  
>> or
>> something else?
>
> Have you tested your scipy atlas ? 2 hours sounds awfully long, and  
> this
> may be due to some bugs caused during a wrong build (for example,  
> errors
> in the masrhalling of arguments at the C/Fortran interface),
>
> cheers,
>
> David
> _______________________________________________
> SciPy-user mailing list
> SciPy-user@scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

***************************************************
Chris Lee
Laser Physics and Nonlinear Optics Group
MESA+ Research Institute for Nanotechnology
University of Twente
Phone: ++31 (0)53 489 3968
fax: ++31 (0)53 489 1102
***************************************************





More information about the SciPy-user mailing list