Isit worth waiting for the launch of PCI-E 3 on motherboards before building a new rig? Also does PCI-E 2 bottleneck the fastest graphics cards on the market.. i.e. GTX 580?
We are not even close to push PCI-E 2.0. The GTX 580 is very well under PCI-E 2.0. PCI-E 3.0 won't provide you with any advantages, at least 4-5 years down the line (this includes having a game that actually PUSHES the GPU to a point that it will need the extra bandwidth). By then, you'll want to upgrade your whole computer in any case.
Also worth noting that neither Sandy Bridge, nor Bulldozer etc, will be able to use PCI-E 3 at first - if we're going off the integrated graphics/PCI-E lanes, they'll still be at 2.0.
But then motherboard manufacturers do like to push out features way before they become native on the chipsets. Would it not be possible to get 3.0 working on the "old" chipsets with an additional chip?
All depends. With P55 stuff, the PCI-E lanes are more or less embedded into the CPU, so in that case, no. I reckon an nForce 200-esque solution could be possible, but probably a bit tricky to implement. There was never anything like that for the transition between PCI-E 1 and 2, IIRC.
Dragonnut, not my knowledge. As you'll need a chip that by pass the north and south bridge and communicate directly with the CPU.. essentially being a motherboard chipset. This is not permitted due to licensing issues. If you do... then you get what Nvidia got as treatment, which is be as vague as possible on if you can or can't make a chipset on the the new CPU, and once you spend several million dollars on it and have it ready for mass production, Intel comes in with a big law suit, just so that you loose even more money with another several million dollars in lawyer fees.
Perhaps you can request a daily/weekly rep service from bit-tech Once you subscribe, it reps a person of your choice once every week or once per day.
So it's safe to assume that with PCI-e 3.0 to be realised next year as a gimmick for companys to make more money, when in actual fact theres no use for it at this current time?
That is something different. Right now it's the race for the fastest SSD possible. Even if Nvidia releases Geforce GTX 600 series, on PCI-E 3.0, and let's say for the sake of argument that it actually reaches the max of the PCI-E 3.0 specs... So what. Name me 1 game that even pushes my GTX 260 under max max max settings at 1920x1200, under 60fps, other than Crysis (actually I haven't tried, I am just assuming it won't run smoothly). So I ask, what's the point? Yea it can go THAT fast, yea you can have Z999999999997 score in 3D Mark, but other than that... what? Game Publisher don't care about the 1-2 million profit games... that's pocket change for them. They want real huge games. And to achieve this, means asking the developers to make games on several platform and/or to be able to run fine on some popular Intel GPU at minimum settings, to ensure to broaden the market, and not make games that pushes our computers, for the sake of awesomeness. Heck we don't even have 64-bit games, except for the 2-3 games out there in total. Sure, unlike the old days, publishers and developers alike, don't take as much risk in the gaming buisness anymore, but who can blame them? They are many great, excellent reviewed games.. but just because it's not a FPS or have is associated with a popular name, it won't and doesn't sell, except on absolute rare cases.
Ah, my bad. I should have said "mechanical hard drives". I think "hard drive", I think 3.5", platters, available over 250GB, for under £50. SSDs are a different class than "hard drives" to me.
Actually my fault... I read "SATA 6GB/s drives", and not "SATA 6GB/s hard drives". I though you were referring to SSD's
