[vtk-developers] Fitting code and it's place in VTK

[Alexandre Guimond]

On Tue, 18 Sep 2001 22:39:17 +0200
Sebastien BARRE <sebastien at barre.nom.fr> writes:

> Same with me. The last time I checked for a NR open source package was
> in november 2000, and I evaluated several free software : MRL, TIM,
> TXMATH, ITL, TNT (at that time I was not aware of VXL). I even had a
> long discussion with one of their authors to see how I could use them
> inside VTK, but none of them was able to fit my requirements :(

I dont think anyone mentionned the GNU Scientific Library
http://sources.redhat.com/gsl/.  I've never used it, but it looks like
it contains quite a bit of good things.  Very NR like it seems.
Written in C and GPLed, but only a beta version at the moment (4th
beta-test release).  Seems to be under pretty active development.

> I was basically looking for an implementation of conjugate gradients
> methods in multidimensions (i.e. minimization) and I ended up
> hacking some code of the Numerical Recipes book :( Slow, not
> portable, etc.

For example, the doc
(http://sources.redhat.com/gsl/ref/gsl-ref_toc.html) has 4 chapters on
minimization and root-finding.  It has at least 2 conjugate gradient
methods. 
-- 
Alexandre Guimond, Ph.D. <guimond at bwh.harvard.edu> Phone: 617.278.0800
http://www.spl.harvard.edu/~guimond/               Fax  : 617.264.5154
Brigham and Women's Hospital, Harvard Medical School
CNI, 221 Longwood Av. RF396, Boston, MA 02115