I am wondering if it makes any sense to seperate a huge table from its indexes by placing both in seperate dbspaces? Can anyone way in on this? I would just like to squeeze every drop of performance I can out of the database side of something before I have to think about changing hardware.

Yes, this makes sense. Especially when the chunks of those dbspaces reside on other disks. Even if this is not the case it could be optimal, because IDS will access it more in parralel.
If possible also consider fragmenting the big tables (and even indexes)across several dbspaces/disks. This can be very nice.

Hope this helps,

Rob Prop