RE: [Algorithms] N-body processing

[Matthew M. <ma...@me...>]

> Matthew,
> Err... I don't think swept volumes (as used in PQP) will help,
> except for 4D
> collision detection which is not what was worrying me :) But maybe I miss
> you point.
>
> Dave,
> Welcome. Stefan actually tried that approach, but later. Well, whatever.
> Could you explain what would be the benefit of lozenges ? Especially a
> lozenge-tree ?!

Dave and I are talking about the same thing here. Swept sphere volumes give
you tighter fits, faster interference rejection, and MUCH faster proximity
queries than OBBs. Particularly important if you're trying to stave off the
poly^2 tests. I assumed from your topic you were talking about N-Body
dynamic simulation, for which fast proximity queries will be extremely
helpful.

(re-)Read the paper "Fast Proximity Queries with Swept Sphere Volumes" at:

You'll be glad you did, unless I'm missing your point. ;^)


>
> /*
> Pierre Terdiman       *   Home: p.t...@wa...
> Software engineer    *   Zappy's Lair:  www.codercorner.com
> */
>
> ----- Original Message -----
> From: Dave Eberly <eb...@ma...>
> To: <gda...@li...>
> Sent: Friday, August 18, 2000 7:56 PM
> Subject: Re: [Algorithms] N-body processing
>
>
> > From: "Matthew MacLaurin" <ma...@me...>
> > To: <gda...@li...>
> > Sent: Friday, August 18, 2000 1:02 PM
> > Subject: RE: [Algorithms] N-body processing
> >
> > > Hey! Check out the swept sphere stuff at UNC! It's the newest
> thing I've
> > > seen in a while for fast proximities and N-Body sim. Even more
> interesting
> > > is what they're using it for -- it appears ideally suited to huge poly
> > > counts. I posted the link last week.
> >
> > I put swept sphere stuff into NetImmerse when UNC was
> > starting to look into the topic as an alternative to using
> > OBBs in a tree.  I had asked Stefan Gottschalk at his
> > dissertation defense if he had considered such bounding
> > volumes, but he indicated 'no'.  They work out quite well
> > because they rely on squared-distance calculations that
> > turn out to be less expensive (on average) to calculate
> > than the separating axes, especially so in situations of
> > close proximity.
> >
> > Code for building trees of spheres (points a specified
> > distance from a point), capsules (points a specified
> > distance from a line segment), and lozenges (points a
> > specified distance from a rectangle) is at my site, the
> > MgcContainment.html page.   I eventually plan on
> > putting the code at my site for the collision detection,
> > but it may take a couple of months.
> >
> > --
> > Dave Eberly
> > eb...@ma...
> > http://www.magic-software.com
> >
> >
> >
> > _______________________________________________
> > GDAlgorithms-list mailing list
> > GDA...@li...
> > http://lists.sourceforge.net/mailman/listinfo/gdalgorithms-list
>
>
> _______________________________________________
> GDAlgorithms-list mailing list
> GDA...@li...
> http://lists.sourceforge.net/mailman/listinfo/gdalgorithms-list
>