Wait time to avoid overloading CPU

[zoligyenge]

Dear Community

What wait time would you use in continuous loops to avoid overloading the CPU?
100ms, 10ms?

For example when checking continuously for collisions or checking motor positions, what freqeuncy is not overloading the CPU? Tech specs state the motors update sensor position at 100Hz, whcih would mean anything lower than wait (10ms) would not make sense, but is that too low for the CPU to handle?

Thanks

[dlech]

On EV3, you need a wait in loops so that the Linux operating system has time to run in the background.

On Powered Up hubs, there is no operating system, so we have the full CPU available. So it doesn't really hurt it to not have a time delay. The difference in power usage is negligible. But if you are using async/await programming then loops do need at least a await wait(0) to allow other tasks to run.

The CPU timer resolution is 1 millisecond on Powered Up hubs, so a 10 ms wait time will work as expected and is a perfectly reasonable amount of time to wait in most control loops.

[zoligyenge]

Thanks, How about on a prime hub?

On EV3, you need a wait in loops so that the Linux operating system has time to run in the background.

On Powered Up hubs, there is no operating system, so we have the full CPU available. So it doesn't really hurt it to not have a time delay. The difference in power usage is negligible. But if you are using async/await programming then loops do need at least a await wait(0) to allow other tasks to run.

The CPU timer resolution is 1 millisecond on Powered Up hubs, so a 10 ms wait time will work as expected and is a perfectly reasonable amount of time to wait in most control loops.

[dlech]

SPIKE Prime is in the "Powered Up" category (along with everything else that has the same Powered Up connector for motors and sensors).