Running CPU-bound tasks

The totalSales() API that we implemented in the Asynchronous request batching and caching section was (intentionally) expensive in terms of resources and took a few hundred milliseconds to run. Nonetheless, invoking the totalSales() function did not affect the ability of the application to process concurrent incoming requests. What we learned about the event loop in Chapter 1, The Node.js Platform, should explain this behavior: invoking an asynchronous operation always causes the stack to unwind back to the event loop, leaving it free to handle other requests.

But what happens when we run a synchronous task that takes a long time to complete and that never gives back the control to the event loop until it has finished? This kind of task is also known as CPU-bound, because its main characteristic is that it is heavy on CPU utilization rather than being heavy on I/O operations.

Let's work immediately on an example to see how these...