[ODE] ODE's Jacobian matrices

Joakim Eriksson jme at snowcode.com
Wed Oct 1 08:52:25 MST 2003


That approximation does exist in the constraint system also
and I doubt that you will find any books on that subject.
It's probably more of an observation that Russ made and 
an approximation he thought would work in the sake of speed.
Getting rid of any sin and cos do impact performance a bit
because they are very-very slow and usually any books or
papers on the subject of physics generally dont talk about
making a simulation worse in the sake of speed.

/Joakim E. - http://www.snowcode.com

> -----Original Message-----
> From: Gary R. Van Sickle [mailto:g.r.vansickle at worldnet.att.net]
> Sent: den 1 oktober 2003 06:50
> To: ode at q12.org
> Subject: RE: [ODE] ODE's Jacobian matrices
> 
> 
> > Gary R. Van Sickle wrote:
> > > IIRC, this has something to do with sin(x) ~= x for
> > > small |x|.  When x isn't small, craziness ensues.
> >
> > In that case, finding a jacobian with dC(q)/dt is an
> > approximation.  Gary, where'd you hear about this?
> 
> I read something to this effect on this list.  Now that I 
> think about it though,
> this was in the context of *un*constrained spinning objects 
> gaining energy and
> eventually "exploding", so it's probably quaternion 
> integration rather than
> constraints at issue.
> 
> Sorry for the confusion, it was late. ;-)



More information about the ODE mailing list