Quote - To me it's more likely some random element of the simulation just required more math to be done

There most likely is a part of randomness involved, but the simulation itself seems to be pretty deterministic, as running it several times with the same settings will apparently yield the same result (as I have discovered to my disappointment while hoping to get nicer creases). So if there is a part of randomness, it's fairly small and shouldn't IMHO make any big difference in calculation time.

The method to know for sure if display has some influence would be to run a simulation with everything made invisible (including the ground plane, which for some reason does eat resources)...   If anyone is bored...  :-D