New Post: Math.Net Numerics Svd vs. DGESVD

The U matrix is as close as the VT matrix if you look at the absolute values, but the first column of the U matrix is negated.

Similar is the VT matrix but here it is the second row which is negated. As I know now, that Math.NET uses LINPACK SVD instead of LAPACK's DGESVD routine I'll try to find something with google. I thought Math.NET uses LAPACK's DGESVD and was wondering why the results are different..

Thanks for that Info!