Is it possible to use atomic operations for thread synchronization without introducing locks?

Yes, atomic operations allow synchronizing access to shared data between threads without using traditional locks (mutex, synchronized). They ensure that the operation is performed as an indivisible unit, preventing race conditions.

For example, in Android (Java/Kotlin), you can use classes from the java.util.concurrent.atomic package, such as AtomicInteger:

AtomicInteger counter = new AtomicInteger(0);
counter.incrementAndGet(); // atomic increment of the value

This helps avoid overhead and issues associated with locking, but it is not suitable for all scenarios — complex operations involving multiple variables usually require more sophisticated synchronization.