each Woodcrest socket will have its own fsb, what's to consider ? Nothing is changing except the cores per socket and having more cache will limit requests to local and non-local memory taking pressure of bandwidth. Sure Opteron will have better 4-core cross latency and for some applications maybe more important but I expect Woodcrest to do very well in the dual dual-core die space. AMD will have to keep pace with binsplit yields to stay up. Obviously more capacity will mean more opportunities for AMD to take share where it couldn't before so it's not really a black and white situation but a complex scenario developing which is not really possible to analyse fully until all the rubber hits the road from both sides.
Market Data 
Markets 
AI
