I was wondering, if I were to put 3 GTX 570s in the Rampage III formula, the PCIe 2.0 bandwidth would drop to 8x on all lanes.
From what I've read, that means a bandwidth 8000MB/s per slot (http://www.directron.com/expressguide.html).
The thing is, what is the bandwidth these cards are capable of, and would I see a bottleneck in terms of speed (between the lane and the mobo), compared to a single GTX 570 that would run at 16000MB/s (16x) ?
Would I absolutly need a motherboard that has 2 NF200 chips (like the EVGA Classified 4-way SLI) in order to utilise the full potential of these cards ?
I think that in theory, the answer is yes, but in practice, does it make a diffrence in games and in benchmarks if the cards are running at 8x or 16x ?
And is it worth it (ratio performance update vs price), because these chips are really expensive, thus making the motherboard very expensive ?
Were a long way from saturating the PCIE 2.0 bandwidth for video cards. Check out the PCIE scaling article. http://www.techpowerup.com/reviews/NVIDIA/GTX_480_PCI-E...
The up and coming PCIE 3.0 is useless at this point in time IMO. Unless your using a OCZ RevoDrive ssd. or some other SSD that takes a PCIE slot.