Re: [Getfem-users] collaboration between GMM and KDE/Eigen

[Yves Renard]

Le Dimanche 28 Janvier 2007 21:49, vous avez écrit :
> Hi Yves,
>    I corrected the test. Stupid mistake. But I'm still
> unclear about what gmm::mult(M1, V2, V2, V1) is
> supposed to do. Or what gmm::mult_add(M1, V1, V2) is
> supposed to do
>
> Isn't gmm::mult_add supposed to do an accelerated
> operation
> M1*V2 + V1 -> V1 st. size(V2) doesn't have to equal
> size(V1)?
Indeed, yes.
>
> Should I just do mult(M1, V2, V1); add(V1, V3, V3)?

You are right. There was a real bug for non squared csr matrices.

lines 750-753 of gmm_matrix.h should be

    static const_row_iterator row_begin(const this_type &m)
    { return const_row_iterator(m.pr, m.ir, m.jc, m.nc, m.pr); }
    static const_row_iterator row_end(const this_type &m)
    { return const_row_iterator(m.pr, m.ir, m.jc + m.nr, m.nc, m.pr); }
 
instead of

    static const_row_iterator row_begin(const this_type &m)
    { return const_row_iterator(m.pr, m.ir, m.jc, m.nr, m.pr); }
    static const_row_iterator row_end(const this_type &m)
    { return const_row_iterator(m.pr, m.ir, m.jc + m.nc, m.nr, m.pr); }

Thank you for this bug report. It should work correctly now. I commited the 
changes at gna.org.

-- 
Yves.

-------------------------------------------------------------------------
  Yves Renard (address@hidden)       tel : (33) 04.72.43.80.11
  Pole de Mathematiques, INSA de Lyon          fax : (33) 04.72.43.85.29
  Institut Camille Jordan - CNRS UMR 5208
  20, rue Albert Einstein
  69621 Villeurbanne Cedex, FRANCE
  http://math.univ-lyon1.fr/~renard
-------------------------------------------------------------------------