This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: LAPACK vs. GSL matrix algebra
- From: <kaska at yatsen dot cas dot McMaster dot CA>
- To: Linas Vepstas <linas at austin dot ibm dot com>
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Mon, 27 Sep 2004 12:35:36 -0400 (EDT)
- Subject: Re: LAPACK vs. GSL matrix algebra
you can easily access the fortran routines from C.
i always write C routines that access the fortran
functions from LAPACK that i need. as long as you
keep in mind that fortran stores matrices column-wise
and C does it row-wise. but most fortran routines
allow you to set the 'transpose' option so you can
easily pass in your C matrix.
that way you don't have to write your own matrix
routines. the fortran lapack is pretty good and it
has been extensively tested over the years.
if you need example files, let me know.
k
On Mon, 27 Sep 2004, Linas Vepstas wrote:
>
>
> Hi,
>
> What is the current thinking with adding to the matrix functionality
> in GSL? I need to find the eigenvectors of a non-symmetric real
> matrix, which is something that GSL doesn't currently support.
>
> My choices seem to be
> -- try to remember how to code in fortran
> -- try to implement something in C, with an eye towards submitting
> it to GSL.
>
> Any suggestions, words of encouragement to go one way or the other?
>
>
> --linas
>