[SciPy-User] solving large sparse linear system with Laplacian matrix

[David Warde-Farley]

On 30-Oct-09, at 4:25 AM, josef.pktd at gmail.com wrote:

> I don't know about memory consumption but I think
> scipy.sparse.linalg.factorized(A)
> should be more efficient if you need to solve for several right hand  
> sides.

You may be right but the factorized matrices may be an awful lot  
denser which could impact her memory requirements.

You may have some luck with lgmres and/or bicg, as far as I know  
neither imposes conditions on positive definiteness.

You might also look at http://code.google.com/p/pyamg/ - they have a  
lot of other methods implemented, possibly some specially optimized  
for symmetric banded matrices.

David