Evaluating Spin Locks
Let's evaluate the spin locks under the metrics of correctness, fairness, and performance.
We'll cover the following
Given the basic spin lock, it can now be evaluated on how effective it is along our previously described axes.
Correctness
The most important aspect of a lock is correctness: does it provide mutual exclusion? The answer here is yes: the spin lock only allows a single thread to enter the critical section at a time. Thus, we have a correct lock.
Fairness
The next axis is fairness. How fair is a spin lock to a waiting thread? Can you guarantee that a waiting thread will ever enter the critical section? The answer here, unfortunately, is bad news: spin locks don’t provide any fairness guarantees. Indeed, a thread spinning may spin forever, under contention. Simple spin locks (as discussed thus far) are not fair and may lead to starvation.
Performance
The final axis is performance. What are the costs of using a spin lock? To analyze this more carefully, we suggest thinking about a few different cases. In the first, imagine threads competing for the lock on a single processor; in the second, consider threads spread out across many CPUs.
On a single CPU
For spin locks, in the single CPU case, performance overheads can be quite painful; imagine the case where the thread holding the lock is preempted within a critical section. The scheduler might then run every other thread (imagine there are − 1 others), each of which tries to acquire the lock. In this case, each of those threads will spin for the duration of a time slice before giving up the CPU, a waste of CPU cycles.
On multiple CPUs
However, on multiple CPUs, spin locks work reasonably well (if the number of threads roughly equals the number of CPUs). The thinking goes as follows: imagine Thread A on CPU 1 and Thread B on CPU 2, both contending for a lock. If Thread A (CPU 1) grabs the lock, and then Thread B tries to, B will spin (on CPU 2). However, presumably, the critical section is short, and thus soon the lock becomes available, and is acquired by Thread B. Spinning to wait for a lock held on another processor doesn’t waste many cycles in this case, and thus can be effective.
Get hands-on with 1400+ tech skills courses.