actually, if an application is not programmed to take advantage of having multiple threads, even os x cannot distribute the task to a different processor. though if there is 2 non-aware apps running, the os can farm the two threads off on different processors, making things quicker. basically, a thread is a task, MP aware apps use many threads, allowing the OS to farm these threads out to whatever processor is more idle. in non-aware apps, the OS only receives one thread, thus it can only use one processor for that app (though it can still switch that thread to the other processor) (think of a manager farming tasks out to multiple workers-MP aware v. MP non-aware would be a 10 page project broken down into 5 2 page sections v. a 10 page project stapled together as one whole section. i always take tests with a grain of salt (my real world experience is all i worry about) but the following shows a DP mac performing tasks in roughly 1/2 the time it took a SP mac of the same speed (for MP aware apps) and non-MP aware apps running just slightly quicker. i dont think a 2x improvement in speed is normal or should be expected, but a mature SMP OS should be able to get 70-80% of the theoretical "doubling" of processor power with no problem. Heck, the VT G5 cluster gets around 60% of the theoretical output, and they have 2200 processors (think of the over head on that) http://barefeats.com/pm1000.html sandor On Feb 7, 2004, at 6:53 PM, Alex wrote: > >> If the application is also dual processor aware you are even better >> off. > > Under OS X it is not necessary for an application to be "aware of" > (actually, "written for") the dual processors to benefit from their > presence. However, it is quite true that an application must be native > and multithreaded to benefit fully.