What a Thread Switch Actually Costs

A context switch happens when the CPU stops running one thread and starts running another. The kernel saves the current thread’s execution state, restores the next thread’s state, and hands over the CPU.

But the more important question is:

How much performance do we lose when the CPU switches threads?

There is no fixed answer. The switch itself requires some kernel work, but the larger cost often comes afterward. The new thread may need to reload caches, rebuild CPU-local state, wait for scheduling, or even migrate to a different CPU core.

So it is better to think of context-switch cost as:

The actual cost depends on the workload, processor, operating system, and how the switch is measured.

This chapter separates direct switch work from cache, migration, scheduling, and handoff costs, compares thread and process switches, and develops a practical method for measuring and interpreting switch behavior on Linux.