Christoph Hormann wrote:
> I have not yet seen any examples on how it performs memory wise (which
> will obviously be *the* problem of such a patch) same for how it
> interferes with radiosity and antialiasing.
> 
Ah, and "memory wise": The patch dumps information about the used memory 
to the user when tracing an image. 

It requires to keep the complete image in memory (hence if there is already 
some image storage in Megapov. it could be used to prevent storing the 
same information twice). And it needs the tile heap which is 16 byte per 
tile (on a 32bit box). The max. number of tiles depends on the rendered 
image; expect 256kb for a 640x480 image. 

But I do not see memory consumption as a very large problem since it 
is meant for design phase; you should not use it when doing the final 
high-resolution render. 

Wolfgang

Post a reply to this message