With the current version (8.2.5), it is not possible to spread the simulation over multiple cores. The main reason for this is the architectural design of ED. Currently there is only one sequential event queue. Developments of ED are continuous and our development team is constantly looking for ways to improve the simulation speed and\or the size of the simulation models. Multicore and multithreading are techniques that have our close attention.
For now, there are several tips we can give to maximize performance. Of course, the hardware plays an important part here. For processors, the clock rate is an important characteristic. But within ED, you can do several things to improve the performance. For example closing all 2D and 3D windows will improve your simulation speed. But also the way you program can improve this. Have a look at this topic: order-of-fast-atom-referencing-t49.html.
Perhaps you can give some figures (size of model, maximum number of objects in your model and actual vs. simulation time\length) of your simulation?