Alex Schultz wrote: > Alex Schultz wrote: > >> I don't see how it could be that bad either: I'm testing firing >> missile swarm (about the most projectile intensive spell there is), >> into a tight 3 director loop, and my server barely gets up to 2% cpu >> usage. Large fireball fired as rapidly as can be by a signal player's >> natural casting can't push it over 1.3% cpu, and even with all this, >> the idle cpu usage is 1%. Really, unless lots of monsters are casting, >> or if it's meteor swarm with it's massive amount of fire, I don't see >> how this could be a problem. > > > And one other curious statistic to add: > I tested level 12 MS with that too, and no matter how much I tried I > couldn't get cpu usage over 13%, However then I tried with level 112 and > that brought it up to 100% cpu, due to the increase in number of meteors > with level. > Overall from what I can tell, directors have little effect on all this > other than that they prevent if from hitting a exploding/disappearing > for longer, which could be simulated anyways by casting out into the > open in bigworld. This was changed in the spell redo - spells now have a range before they run out. So loops aren't quite as bad. However, have two directors pointing to each other will still be much worse than a player firing those same missiles on the bigworld. That is because whenever one of these objects move onto a space, it has to check all the other objects on the space to see if anything happens. So if you fire across the world maps, there will only be a few objects on each space the meteor has to check against. In comparison, in the director loop, it has to check on all the objects on the space, which is considerable (all the other meteors). And you have a whoel bunch, so this is happening a whole lot. So at some level, it isn't any worse if there is only one meteor going back and forth . But if you have 100 going back and forth, that hundreds of times worse than if those were being fired across an open map.