Ghost iteration statistics

This MR adds some functionality that I created a while ago to analyse the smoothing length iterations for hydro, stars and black holes. It adds some counters to every cell to keep track of the number of retries per iteration and the average value and standard deviation of the smoothing length for every iteration. The number of iterations to track is a configuration parameter.

The new script that goes with this creates a plot like this (illustrated on a hydro only low z EAGLE 6, for lack of better IC):

This was somewhat helpful when I was trying to optimise the COLIBRE code, and could still be useful if we ever decide to change the way we determine smoothing lengths for stars and black holes.