At Sat, 6 Aug 2011 01:08:06 +0200,
Sazonov, Vasily ([email protected]) wrote:
> I suppose, I found a bug in 'gsl_eigen_symmv' function. I have examined it on 
> matrix:
> 
> 7.255197 0.000000 
> -2.666667 17.177062 
> 

I noticed that your matrix is not symmetrical, so gsl_eigen_symmv is
not going to give the result you expect - it assumes a symmetrical
matrix.  

The lower triangular part is used, so you are actually computing the
eigenvalues of [7.255197, -2.666667; -2.666667, 17.177062], and the
correct results for that are 6.5839, 17.8484 as given.

-- 
Brian Gough


_______________________________________________
Bug-gsl mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-gsl

Reply via email to