|
Ublas : |
From: Gunter Winkler (gunter.winkler_at_[hidden])
Date: 2006-10-04 02:50:38
On Wednesday 04 October 2006 04:32, Nico Galoppo wrote:
> Gunther
>
> I know this thread is not recent, but as a follow-up, does your code also
> work for sparse matrices with
>
> typedef boost::numeric::ublas::compressed_matrix<TINY>
> BIG;
The code fails, because TINY is not comparable with zero (or any other real
value). This could only be avoided if we had some scalar_traits<T>::zero
class that is used in all algorithms.
mfg
Gunter