Mike Williams wrote:
> 
> I'd guess that the vast majority of the tracing time would be consumed
> by looking for the bounding boxes which have been automatically
> generated. Each ray has to be checked against all 1000 bounding boxes,
> but once the bounding boxes have been located, the ray only has to be
> checked against at most one actual object. (Because the object exactly
> fills the bounding box, if there's no reflection or transparency the ray
> will always terminate at the first object). So the construction details
> of your object would only be expected to have a tiny influence the trace
> time.
> 

your argumentation sounds quite reasonable, even though objects are
partly transparent in this case :-) I also looked more on the memory 
point, where differences are much larger.  

Christoph

--
Christoph Hormann <chr### [at] gmxde>
Homepage: http://www.schunter.etc.tu-bs.de/~chris/

Post a reply to this message