Wire loops are slow

[TodePond]

When working on bigger simulations, a long wire loop can cause stuttering because of trying to figure out if it's a loop in getpeak

Ideas:

• reorder cache tree for quicker miss detection.
• don't clear the cache so eagerly (would this stop it doing pointless extra cache lookups?)
• change getpeak to iterative
• figure out how to figure out loops without just brrrrrr-ing it
• other ideas?

[TodePond]

Maybe partly improved by #167 ?

[TodePond]

Definitely improved, but I've seen it being slow when adding wires to a 'semi-pulsing loop'.

• big call stack (though less big than before).
• lots of objectEquals. <<<<< That's gotta be something to avoid, surely. Can we do some sort of id'd system?

Or one of the previously mentioned solutions

[TodePond]

Rewrote the Nogan cloning function - should be a bit faster, but doesn't solve the base problem

I think I'm gonna have to rewrite getPeak as imperative. Big task, so downgrading this to stretch goal