Christoph Hormann <chr### [at] gmxde> wrote in
news:3CF### [at] gmxde: 

> 
> I tried some larger grids of masses with my simulation system.  It of
> course takes much longer, several hours on an Athlon 1 GHz.

Hmm, my 5x5x5 piece renders significantly faster. But it does look good.

I believe the cause of this might be the way your system handles 
connections, but I can't be sure until I see the source. :)

--
Cheers,
Gergely

Post a reply to this message

> #10 won't look so good with a mesh. If you look closely you'll see
> that the most contorted part of the cube becomes inverted (inside-
> out).

Hm. You're right. But if we knew where the outline/border of the box is, it
could be used for self collision detection.

Post a reply to this message

Gergely Vandor wrote:
> 
> [...]
> 
> Hmm, my 5x5x5 piece renders significantly faster. But it does look good.

Thanks.

> I believe the cause of this might be the way your system handles
> connections, but I can't be sure until I see the source. :)

The main influence factor is the number of integration steps.  I used 30
per frame with a Heun integration method (i.e. 2 evaluations per step).

Christoph

-- 
POV-Ray tutorials, IsoWood include,                 
TransSkin and more: http://www.tu-bs.de/~y0013390/  
Last updated 02 Jun. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message