Re: [ublas] Simple problem with LU-decomposition.

From: Paul C. Leopardi (paul.leopardi_at_[hidden])
Date: 2008-08-25 20:48:55

On Tue, 26 Aug 2008, Gunter Winkler wrote:
> Иволгин Ардалион schrieb:
> > Check failed in .../include/boost/numeric/ublas/lu.hpp at line 157:
> > singular != 0 || detail::expression_type_check (prod
> > (triangular_adaptor<matrix_type, unit_lower> (m),
> > triangular_adaptor<matrix_type, upper> (m)), cm)
>
> This assertion say that your matrix is either singular (which is true
> for the given example) or the product of the two factors differs from
> the original matrix (which is also true if the factorization broke down)
>
> It should work for a non-singular matrix.

Hi Gunter,
Sorry to jump in at this late stage, but am I correct in assuming that this 
assertion is only triggered in DEBUG mode? GluCat uses uBLAS for LU 
decomposition on matrices which are not known to be non-sngular, and uses NAN 
to cover the cases where singularity is detected.
Best, Paul