jimbobjim wrote:
> 
> [...]
> 
> Although I have a question about what pov does during the parse stage.
> During test renders I was using ~30,000 objects (pyramid2 from shapes.inc)
> for which the total trace time was just a few minutes. For the final render
> I used ~60,000 of these objects and even though the trace only took a few
> minutes, pov took 25 minutes to during the parse to create light and vista
> buffers. Is there anything I can do to speed this process up? I changed the
> union to a mesh in the pyramid2 definition and that helped loads - peak
> memory used dropped from 50Mb to about 14Mb in the 30K object case - but on
> the final render peak memory was 125Mb and my machine has only 64Mb so
> windows was thrashing the virtual mem a bit. (maybe I just answered my own
> question)

First of all you must have got something wrong here.  'Pyramid2' is in
'shapes2.inc' and is an intersection of planes which is surely much slower
and needs more memory than a mesh.  

The best solution would probably be to use meshes placed in a tree
structure (can be achieved using recursive macros)

If you want to optimize render speed it would be even better to make all
pyramids in one mesh, but this would result in a higher memory use and
parsing time so it's probably not a practicable idea.

Christoph

-- 
POV-Ray tutorials, IsoWood include,                 
TransSkin and more: http://www.tu-bs.de/~y0013390/  
Last updated 21 Feb. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message