So I should go for a X79 platform with a GTX 680? Sorry, I am a bit slow atm.
Yes, I see what you mean. Start to worry about future technolgy that hasn't been really implemented yet. Thanks for the answer! +1
I would say you don't need x79 for 1 680, but I say if you can afford it already and you are gonna do other stuff with your computer than just game and serch the internet, the x79 is the most flexable platform out, it can do anything you ask of it now, and will continue too longer into the future.
z68/z77 boards will do great with 1-680, with 2-they will also do fine, you could argue that each card is giving you 100% of what it could, because hey are bacically sharing resorces, 8x each. This would only be a few % at most and that is what PCI-e 3.0 is trying to fix for the lower end platforms. 8 3.0 lanes running faster can = 16 2.0 lanes running slower.
The problem is games don't really need the speed of 2.0 yet, and the grapfhic cards/ chipsets can't really tranfer data at 3.0 speeds yet, but they are working on it. With the absents of high speed transfer to get more data thru on less lanes I feel its better for now to have more slower lanes. This is what the x79 platform has, the abillity to run 2 cards at full 2.0 16x(lanes).
With one card your covered with both platforms, with the z77 having the slight advantage if you wait for the IB CPU to be released, but without thoose they serve up the same data to your GPU at the same speed. The differance comes in the future when you add another 680 for SLI.
the x79 still has another 16x 2.0 slot so possable data tranfer doubles with the need for it to be doubled with the addition of the second card.
z77/z68 with a Sandy Bridge CPU deals with it by having to cut the avalible data in half, where each card gets 8x at 2.0 spped. Is sounds bad, and like something you really don't want to buy into when buying a new machine but the reality is it dosn't really cause a problem.
the same platform with an IVY Bridge CPU can run twice as fast in getting the data out, which is useless with 16x(lanes) because no graphics cards are useing 2.0 speeds yet. but when you add in the second card, it requires the data transfer(bandwith) to be cut in half, OK no problem, It was already twice as high as it needed to be in the first place.
so really as it stands to day, we are all planing out our future rigs to not be bottleneceked in a place where there never really was a bottleneck ever in the past, as PCI-e specs have stayed ahead of the curve and develpoed before the hardware needed it. If the don't develope PCI-e 3.0 standards before we need it , how can hardware be made to use it??? If the hardware needed PCI-e 4.0 speeds already, why would they be wasting time with silly 3.0 speeds???
My current system shouldn't even be playing current modern games, all of them recomend a quad core processer and a minimim of a low end PCI-e 2.0 graphics card. Well I got a dual core processer(several generations old), a mid range(current generation) 2.0 grapfics card, but running it on a PCI-e 1.0 16x Motherboard, and it runs all of them good enough.
Do I think there is a bottle neck there? YES I do!!! Is it the CPU?? NOPE!! none of the games max out both the cores. Is it the Grapfics card?? Nope!! the same games on different machines with this card get higher frame rates.. So what is it?? I would have to say it is the PCI-e 1.0!! How much it lacking?? I would say 10-20%
What have I learned fron my old machine?? PCI-e 1.0 at 16x is only about 10-20% to slow to feed a mid range 560Ti PCI-e 2.0 card In most current games.
