I'd love to find out more about your 3D sound engine for PD. Is the source code available?
No, nothing has been made public yet. I have cooked up a few externals to deal with 4, 5.1 and 8 speaker arrays (8 in a cube). They also deal with distance calculation and "camera" rotation around the 3 axis using OpenGL terminology. Then there is as c++ stub that eventually has to be included into vr++. It takes care of net connection, voice allocation/management, formatting net messages for PD and other stuff that will limit the stress on PD.
I am starting work on a PD cluster with the idea of using many computers to generate a 3D array of speakers with the advantage of spreading out processing as well.
In my situation the sound server will run on a single machine in either a single/multiprocessor or cluster environment. I have made no attempts to split the sound process out on more machines. I fear that the precision I requre will not be possible.
If you are looking for at multi processor / multi threading programming framework vr++ may be of interest. It can be used for other things than VR aswell :-) (You should contact Henrik R. Nagel - his address is on the web page somewhere).
Cheers Soeren