[ODE] Multi-threading

Jon Watte (ODE) hplus-ode at mindcontrol.org
Fri Aug 19 13:50:44 MST 2005


> It seems to me that most of these arguments (i.e. referring to "finer
> control") are referring to functional parallelism, though, not data
> parallelism (which I think is the case with parallel physics/collision
> detection).  As long as the problem can be broken into a bunch of
> small chunks, each being processed in exactly the same way, we have
> data parallelism, and that's what OpenMP is good for.  I don't know

Assuming you can keep the data you need in cache in each of the CPUs, so 
they don't step on each others toes trying to get at the same main data 
store in RAM. And assuming that the threading overhead of the data 
parallelism doesn't dwarf any gains.

My prediction: Next-generation multi-core consoles and PCs will do much 
better with functional parallel approaches than "macro data parallel" 
approaches, for games. ("macro data parallel" to separate it from SIMD 
processing, which is a good idea when you can do it)

Cheers,

			/ h+


More information about the ODE mailing list