openmw.org

I had an idea.

How about detecting the number of cores of the CPU in engine.cpp and processing each call to frame() in a different core to improve the performance?

It's a good idea? Or isn't possible because dependencies between frames?

I will try to compile the code, I'm starting with it. Noob.

Rendering is done in the graphics processing unit. Software Rendering (rendering in the CPU) is not widely used due to performance drawbacks, CPU's are simply not designed for this kind of data processing.

Ohh of course, i see. My mistake. Thanks.
What areas could be threaded to speed up? Physics, sound, ai?

Im no dev or pro, but i think AI or physics should have some attention.