🚀 go-pugleaf

In article <Bz5rC9.1A9@rice.edu> steve@ion.rice.edu writes:
>>
>>For the last few days I've been working on a molecule display app using  
>>renderman. It's now working perfectly, with one exception. Whenever I try to  
>>draw a large molecule (300-400 spheres), renderman chokes. I get a message on  
>>the console about renderman losing it's connection to the window server or  
>>something. After digging around the filesystem for a while, I located the place  
>>where the renderman daemon writes ITS errors, and discovered a variety of  
>>messages all boiling down to "out of memory", including some that explain why  
>>it can't connect to the window server (memory problems again). Thinking back, I  
>>remember similar problems when I tried to put too many spheres in when  
>>debugging Plot3D.   Is there any way to increase the amount of memory the  
>>renderman daemon has available to it???  NeXT's complete lack of documentation  
>>doesn't help the situation at all!    Any advice at all would be greatly  
>>appreciated!



Well, there's documentation, and then there's documentation.  I'm staring
at some mighty fine 3.0 paper doc right now, but unfortunately, not everything
has a man page...

Anyway, given the relative fragility of qrman (compared to prman), I'm not
surprised it's choking.  I can't believe there are any compiled in limits
for the amount of geometry, so you might try this on a machine that has
a good amount of memory and lots of swap space and see if it dies in the 
same way (i.e. same size database).  I'm killing qrman pretty repeatedly
these days, when I give it badly shaped polygons, and it wedges it well
enough make it that I have to reboot.  We can only hope 3.1 has a more
robust qrman. 

-- 

-->  Michael B. Johnson
-->  MIT Media Lab      --  Computer Graphics & Animation Group
-->  (617) 253-0663     --  wave@media-lab.media.mit.edu