We do get minor usage of the other CPU thread; it appears that some of the kernel stuff may actually be using the other CPU for its work while the main CS thread is doing its stuff. In other words, a dual-cpu system appears to run a wee bit faster than the equivalent single-cpu system; I\'ve been able to test this by disabling one of my CPUs using the CHUD tools.
Beyond that, though, no - there is no SMP in CS; this is true in the wider scope, in that we don\'t have any background threads processing information, and in the smaller scope - we don\'t have any parallel algorithms. In addition, there is currently no AltiVec support in CS; that will change, but not for MolecularBlue release this week.
Having said that, the more we hand off to GL, the better off we are in terms of both AltiVec and MP support, where Apple is, behind the scenes, free to make use of whatever resources are available to parallelize and vectorize any necessary work. This will, in future releases *but not in this one* mean that we get much better performance than we are today under 10.2.
We\'ve also run into a set of issues under Jaguar regarding the OpenGL implementation that I hope might be different in future releases; I\'m told whispers that certain GL changes in 10.3 may be of benefit to our cause.
Keep your fingers crossed.
