Re: [ublas] bug on compressed_matrix ?

Subject: Re: [ublas] bug on compressed_matrix ?
From: Riccardo Rossi (rrossi_at_[hidden])
Date: 2013-01-16 14:48:22

mmm

is there any clean way of constructing a ublas matrix given the 3 csr
vectors?

riccardo

On Wed, Jan 16, 2013 at 8:20 PM, Thomas Klimpel <Thomas.Klimpel_at_[hidden]
> wrote:

> Hoang Giang Bui wrote:
> > Yes, it does the trick. And I can handle the index2_data and value_data
> > without problem if index1_data is correct. Thank you very much. But this
> > confusion could be avoided if it is mentioned somewhere in the
> > documentation.
>
> You are right. At least the documentation of the sparse stuff seems
> incomplete and misleading to me. I think David Bellot started a new doxygen
> based ublas documentation at <ublas.sourceforge.net>.
>
>
> > Btw, is there any performance penalty if complete_index1_data is called?
>
> No, there is no performance penalty if "complete_index1_data()" is called
> before data is passed to an "external library" or your own code.
>
> Regards,
> Thomas
> _______________________________________________
> ublas mailing list
> ublas_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/ublas
> Sent to: rrossi_at_[hidden]
>