What is CPU affinity, and how does binding a process to a specific CPU or subset of CPUs affect scheduling on multi‑core systems?

• A. CPU affinity is a file permission mechanism for CPU usage.
• B. CPU affinity binds a process to a specific CPU or subset to improve cache locality and reduce context switching.
• C. CPU affinity enables processes to migrate freely between cores without any performance impact.
• D. CPU affinity is a hardware error detection feature in CPUs.

Answer: (B)

CPU affinity is about binding a process (or a thread) to a specific CPU core or to a defined subset of cores. On a multi-core system, the scheduler normally moves work around to balance load, but when a process is fixed to a particular core, the scheduler tries to run that process on that core whenever possible.

This binding improves performance mainly through cache locality. If a process keeps running on the same core, the data it uses stays warm in that core’s caches (L1/L2), so it can access memory faster and with fewer cache misses. It also often reduces the number of times the CPU has to switch context between different processes, which saves the overhead associated with saving and restoring state.

There are trade-offs: pinning a process to one core can lead to underutilization if that core becomes a bottleneck or if other cores remain idle, and it can limit the scheduler’s ability to balance load across all cores. Using an affinity set that includes multiple cores gives some flexibility while still restricting migration, which can be useful for predictable performance or for isolating workloads.

The other options don’t fit: CPU affinity is not a file permission mechanism, it does not allow free, unrestricted migration without performance impact, and it is not a hardware error-detection feature.