BACKGROUND: I will start off my saying that I'm not very experienced with performance tuning using filegroups. I understand how they work, but admittedly am not, for example, separating off my data and nonclustered indexes to separate spindles of drives on my SAN.
I've been reading more and more that this can be a helpful and underused way to increase speed, but I've also read that on larger SANs when you already have a ton of disks, the impact is actually not that big.
I've also been reading that large, often-joined tables can be separated out into different drives via filegroups in order to provide those queries a performance boost, but again, I have no hands on knowledge of how much of an impact this really makes. We're already achieving very good performance here, but of course I'm always looking for ways to improve.
So now with that...