On an original PI Zero, running NR 4.0 on Bullseye, with Python 3 I find that for each gpio pin configured there is an instance of the nrgpio script running. Each one of those consumes about 3% of the processor, even when there is no IO activity. Even on a Pi 3 it is 0.7%.
Looking at the source of nrgpio I deduce that the python input function may be polling stdin which is not ideal if I am right. Perhaps there is no alternative.
I have not been able to try it on python 2 to see if this is worse with 3, as I don't have a pi running 2.
On an original PI Zero, running NR 4.0 on Bullseye, with Python 3 I find that for each gpio pin configured there is an instance of the nrgpio script running. Each one of those consumes about 3% of the processor, even when there is no IO activity. Even on a Pi 3 it is 0.7%.
Looking at the source of nrgpio I deduce that the python
inputfunction may be polling stdin which is not ideal if I am right. Perhaps there is no alternative.I have not been able to try it on python 2 to see if this is worse with 3, as I don't have a pi running 2.