The Pci-e 2.0 slot is more than capable of handleing the GTX 680 and will not restrain it at all , there are currently no cards out that will use up all the bandwith of a Pci-e 2.0 slot. I don't know for certian why the Pci-e 3.0 was relaesed now when there are no cards that can take advantage of the doubling of the bandwith but I guess in the near future there will be so maybe it's best to get the technology out there and working so when the cards are released that will take advantage of it the technolgy will be completely stable and ready to go.
May 5, 2012 4:33:51 PM
Hey izone. Thanks for the quick reply. So the GTX680 doesnt acutally have a performance increase using a 3.0 vs 2.0 slot?
I just don't understand like you said why they would make a 3.0 enabled card and 3.0 motherboards but there is no differnce. I mean if they made 3.0 motherboards but still only 2.0 cards then I could understand the "potential" upgrade and keeping it "future proof" for a few years if building a new system..
Are you aware of any good articles on this topic and possibly benchmarks?
PCIe 3.0 on the cards and motherboard improve the bandwidth problem for multi card setups.
PCIe 3.0 also helps many compute workloads that involve sending and receiving gigabytes over the PCIe bus. The GTX 680 has horrible dual precision compute performance, but it has decent single precision compute performance and PCIe 3.0 might help that.
PCIe 3.0 is also necessary for some extremely fast PCIe solid state drives. OCZ sells two drives that I know of that can saturate a PCIe 2.0 x16 slot. Many servers only have room for x8 slots, so PCIe 3.0 is needed for these cards on such servers. The 12TB model can hit up to 7.2GB/s consistently and the 16TB model can hit up to 6GB/s consistently. A PCIe 2.0 x8 slot can only handle up to 4GB/s.
For gaming, a PCIe 2.0 x16 slot is enough. A PCIe 2.0 x8 slot is a slight bottleneck. This will improve CF/SLI on pcie 3.0 compatible LGA 1155 boards with a PCIe 3.0 compatible CPU (Ivy Bridge). It should also help quad card setups on the Sandy Bridge E platform.
That link has a math problem. The x16 lanes are stated to have 32 times more bandwidth than the single lane connection! That's wrong. A PCIe 1.x lane has 250MB/s. 16 of them have 4GB/s. A single PCIe 2.x lane has 500MB/s. 16 of them have 8GB/s. A single PCIe 3.0 connection has 1GB/s. 16 of them have 16GB/s. However, that article states the x16 bandwidth as double these correct numbers.
Also, I'd like to see a comparison of PCIe 3.0 and 2.x that has Civ 5 in it. That game is one of the few that is fairly new and involves some compute work during play and it might benefit form PCIe 3.0 more than most other games. The same might be true for the Kepler cards when they are using PhysX.