Let's visualize a Linux system with four CPU cores and, for simplicity, one ready-to-run thread. On which CPU core will this thread run? The kernel will decide this; the key thing to realize is that it could run upon any of the four available CPUs!
Can the CPU(s) it could possibly be run upon be specified by the programmer? Yes, indeed; just this feature alone is called CPU affinity. On Linux, it is a per-thread attribute (within the OS). The CPU affinity can be changed on a per-thread basis by changing the thread's CPU affinity mask; this is achieved, of course, via a system call. Let's take a look at the following code:
#define _GNU_SOURCE /* See feature_test_macros(7) */#include <sched.h>int sched_setaffinity(pid_t pid, size_t ...
