This is quite opinion-based. Basically, you may benefit from a RTOS if you have numerous different, complex processes that happen simultaneously. We are talking mid- to high end microcontrollers where you are well beyond 50kb program space and probably beyond 100k LoC. Or when you find yourself constructing your own scheduler abstraction layer, basically a simplistic OS but without separate stacks per process. Where to draw the line exactly is a subjective program design decision.
1 comment thread