I noticed the Flux plugs having a really high timing (166μs per plug), but also that this tends to go up when i add more Flux points. If i remove one of the Flux plugs, the global time stays about the same but goes up a whole lot for the remaining one. So i'm assuming the Flux Plugs actually bear the weight of the whole Flux network when it comes to Opis measuring. Is that correct, or is there some oddity happening with the Flux plugs ?

Also at the same time, my idle Docking station with a Scientific calculator in there clocks at 23μs, it's not bad at all but significantly higher than other Calculator machines (most are between 6 and 15, the various separators, the circuit extracting machines, etc). I thought i'd report that there as you care about the performance data.

The version is still 1.9.9 so, sorry if it's not true anymore, the patch notes didn't mention anything like that though.

I can provide more info on the setup (screenshots, etc) if needed. :)

The Flux system was written a while ago and isn't the most efficient. Yes the plugs basically run the whole system so they will display the biggest times. I'm rewriting this for 1.8.9 systems to be much more efficient, but I doubt many of these features will be ported back, as I want to stop developing on two branches at once.

Alright, thanks ! It makes sense and the timing itself isn't bad for a whole network anyway.

if i can pitch in: i think the best would be to just have the plugs tell the network(object) how much energy they can provide or are requesting, then use a tickhandeler (end of the tick) that makes the network move the energy between the nodes

that way the entire network only has to be traverced twice (1x to determine the flows, 1x to move it all)

Yeah that is what I've done for the 1.8.9 version.