Advanced Python Programming

The downside of locks

In Chapter 19, Deadlocks, we covered an interesting phenomenon, in which the use of locks can lead to undesirable results. Specifically, we found out that, with enough locks implemented in a concurrent program, the whole program can become sequential. Let's analyze this concept with our current program. Consider the Chapter21/example3.py file, as follows:

# ch21/example3.pyimport threadingimport random; random.seed(0)import timedef update(pause_period):    global counter    with count_lock:        current_counter = counter # reading in shared resource        time.sleep(pause_period) # simulating heavy calculations        counter = current_counter + 1 # updating shared resourcepause_periods = [random.randint(0, 1) for i in range(20)]########################################################################### ...