Re: [ublas] About uBLAS solve (Ax=b)

From: Karl Meerbergen (Karl.Meerbergen_at_[hidden])
Date: 2005-01-13 10:51:23

Karl Meerbergen wrote:

> Consulting a book about numerics, you would see that the determinant 
> is not a good measure of singularity. It is better to compute the 
> condition number. There are estimators in LAPACK, functions xTRCON(). 
> But no bindings are provided for these estimators, I think.
>
> Karl 

Actually the function should be xGECON and not xTRCON, what I suggested 
earlier. Sorry about this. xGECON compute an estimate to the inverse of 
the condition number.

The svd-lapack bindings in boost-sandbox can also be used to compute the 
(exact) condition number (largest singular value / smallest singular value).

Karl