One example of putting unrelated threads of work on the same processor would be if one of your threads was handling I/O. Since this thread would have to wait for disk reads (and would be inactive a lot of the time) putting another thread on the processor that had to do a lot of computations would keep the processor in use more of the time.

This comment was marked helpful 0 times.

Question -- it's usually the job of the OS to map threads to processors -- how does that play into / affect these mapping decisions? Is it even possible/tractable for the programmer to explicitly schedule threads, or to specify that a group of threads should be mapped together?

This comment was marked helpful 0 times.

Yes, most systems do provide the programmer mechanisms to set the affinity of a thread. A thread's affinity specifies (or strongly hints at) what physical resource it should be mapped to. For example, if you are interested in mapping a pthread to a particular processor you may be interested in the pthread library function pthread_getaffinity_np().

You can also set the affinity of a process using sched_setaffinity().

This comment was marked helpful 0 times.

You can also reconfigure the scheduler using sched_setscheduler. The SCHED_FIFO scheduler is (roughly) a cooperative scheduler, where your thread will run until you block on IO or call sched_yield (or a higher priority process becomes runnable).

This comment was marked helpful 0 times.