Calculating Condition Numbers

Hello,

I apologize if this is a FAQ -- I searched Google, but didn't find
anything.

My application uses very large, sparse MTL matrices.  What is the "best"
(fastest, relatively precise) way for me to calculate the one-norm
condition number of my sparse matrix?  

Currently, the easiest way that I've found is to copy the matrix into an
LAPACK matrix and run 'gecon' -- but, if I understand the code
correctly, the LAPACK matrix isn't sparse, and with the larger matrices
I'm using, this approach is taking considerable resources.

Has anyone else written code to handle condition numbers with sparse
matrices?


Thanks for any help that you can offer,
Kris

_______________________________________________
This list is archived at http://www.osl.iu.edu/MailArchives/mtl-devel/