As we mentioned previously, due to their communicative and associative properties, reduction operators can have their partial tasks created and processed independently, and this is where concurrency can be applied. To truly understand how a reduction operator utilizes concurrency, let's try implementing a concurrent, multiprocessing reduction operator from scratch—specifically the add operator.
Similar to what we saw in the previous chapter, in this example, we will be using a task queue and a result queue to facilitate our interprocess communication. Specifically, the program will store all of the numbers in the input array in the task queue as individual tasks. As each of our consumers (individual processes) executes, it will call get() on the task queue twice to obtain two task numbers (except for some edge cases where there is no or only one number left in the task queue), add them together, and put the result in the result queue.
Similar to adding pairs...