Generally I do not run code on more than a couple cores unless they have absolutely no dependencies/spin locks since software semaphores/mutexes are essentially a failure for too many cores. So, I generally write data plane apps that perform few functions really fast. So, I'd like to have the ability to limit the SMP core usage and then run a tight forever loop to perform some data I/O or data plane code. Are there any instructions for launching the Linux on a fixed subset of cores and reserve the other(s) for other purposes. Is there a way/demo of loading and running a bare-metal app on specific core(s)?