[SciPy-user] umfpack installation (and dsolve error)

Nils Wagner nwagner@iam.uni-stuttgart...
Mon Mar 23 12:44:04 CDT 2009


On Mon, 23 Mar 2009 17:38:27 +0000 (UTC)
  Eric Friedman <ejf27@cornell.edu> wrote:
> Nils Wagner <nwagner <at> iam.uni-stuttgart.de> writes:
> 
> 
>> > btw, the reason for this is that I'm having a problem 
>> >with dsolve which I think
>> > is a bug, but may just be my error.
>> > 
>> > for example:
>> > 
>> > a=matrix([[1,1],[1,1]])
>> > b= matrix([.5,.5]).T
>> > dsolve(a,b)
>> > yields [.5,.5]
>> > which is not a solution of ax=b  - I think...
>> > (I also entered the matrix as sparse)
>> > 
>> > any ideas what's going on here?
>> > 
>> > thanks,
>> > Eric
>> > 
> 
> 
>> Hi Eric,
> ... 
>> Please not that your (dense) matrix is singular.
>> 
> 
> sorry, I should have mentioned that I need to be able to 
>solve singular matrices
> and will settle for any solution (since there will be 
>many of them).
> (For nonsparse I've been using linalg.lstsq, which works 
>fine for this.)
> 
> any suggestions on which sparse solver to use?
> 
> thanks
> Eric
> 
  
Eric,

Propack might be an option

http://soi.stanford.edu/~rmunk/PROPACK/

AFAIK no python wrapper is available until now.
I have also filed a ticket

http://projects.scipy.org/scipy/ticket/330

Cheers,

          Nils


More information about the SciPy-user mailing list