Sometimes, with infrequently occurring interrupts, the result is actually better performance, because the associated status bit doesn't have to be examined as part of some other event process.

Some years back, I had a project in which there was a display to scan and a very long timer, actually a counter, that interrupted the CPU at a very small number of Hz. Doing the monitoring "manually," i.e. as part of the display scanning, one segment at a time, routine would have been awkward and somewhat inconvenient due to the low system clock frequency. The code for the entire system was smaller than 512 bytes and took less than two hours to generate and debug.

RE