cache each modifier to avoid unnecessary calculation - results in MASSIVELY increased performance and playback.

If you have a large modifier stack and you make a change on one of the lower down modifiers, then all of the modifiers above it recalculate, which is silly, and extremely costly in terms of performance, particularly during timeline playback.

each modifier needs to cache it's own output and only recalculate if one of the above modifiers is modified either directly or indirectly (changing something else in the scene that the modifier is dependant on). caching just the las modifier would be insufficient, because if you changed the last modifier, then all of the above modifiers would need to recalculate again as well.

In situations where this becomes problematic due to deteriorating RAM availability, then the modifier stacks with the lowest calculation time should have all but the last modifiers cache turned off, this frees up memory whilst still not reducing performance (unless changing parameters)