
Ublas : 
From: Preben Hagh Strunge Holm (preben_at_[hidden])
Date: 20070305 09:34:51
> You could use a custom class and overload the prod() function (or simpler
> provide a special axpy_prod) which looks like
>
> project(y, range(r1,r2)) += prod( submatrix, project(x, range(c1,c2)) )
The projectfunction?
Could you explain what should be going on here?
> where submatrix has size (r2r1)x(c2c1). Using a sparse matrix is also
> possible but may be slower because you have to store the column indices.
> Usually the dense algorithm wins over the sparse when there are more than 25%
> nonzeros in you matrix.
>
> Alternatively you could develop a new blocksparse matrix class (which would
> be very interesting) and contribute it to ublas.
I don't think I have the competence in C++ to do that.. But yes  quite
interesting.
I'll look more into the sparse matrix. Sometimes there are only 3 or 6
elements in the matrix of size 72x72 (that's below 25% "I think").
Thanks for helping
Best regards,
Preben