Peformance difference between PCI x16 2.0 versus PCI x16 1.0?

I have only been able to find one HD3850 card that supports a PCI 1.0 slot: http://www.visiontek.com/products/cards/retail/3850_256_PCIe.html

How much slower do you think this card is than a PCI 2.0 sibling?

 

bump...............

 

I doubt there would be any gigantic gains ( maybe a few FPS? maybe? ), PCI-E 1.0 should be enough bandwidth for that card...when you get to like a 3870 and 8800GT or GTS then PCI-E 2.0 would be beneficial but not entirely necessary.

 

I do not think PCI-E 2.0 will provide much better performance when comparing two current cards (such as a 3870) of the same type but with different interfaces (i.e. one is PCI-E 1.0 and one is PCI-E 2.0)

The standards introduced with 2.0 will pay off with cards that are better designed for the standard. Current cards don't seem to really play with the possibilities yet.

Even the 3870 X2 utilizes PCI-E 1.1, due to the interface/bridge chip. It doesn't seem like ATI would've gone with 1.1 if it would've created any significant performance decrease.

 

Actually the 3870 is PCIe 2.0, the internal bridge chip is PCIe 1.1 because they didn't bother with a PCIe 2.0 bridge with such short wire lengths and not chipset jumps.

Anywhoo, for such a card (HD3850) very limited use for PCIe 2.0 and the primary benifit would be improved Hypermemory speed, but overall as mentioned before, the difference would amount to small percentage differences that may or may not translate to a whole frame or two. The smaller the memory size and the greater the need to swap textures likely the more the card will benifit from higher speeds, but overall the difference would be very minor.

Here's a pretty good look by the guys at Tom's Hardware;
http://www.tomshardware.com/2008/01/23/crossfire_meets_pci_express/page5.html

The fps difference is noticeable because the fps number is huge so 2% is a big deal, but 2% on 30fps is only 0.6 frames per second so not even a frame and of course at 60fps it would be just over 1fps improvement.

What would've been nice is if they had looked at the min fps which is likely where you would see chokepoints expressed the most.

 

I don't think people quite understand why current gen cards will NOT benefit at ALL from PCIe2. Once the info is loaded onto the video card, it stays there waiting to be rendered for the duration that it's processing the 3D app.

During actual gameplay, most of what the game needs is already loaded to the video card, including the huge files such as texture data. This means that the ram is only used when streaming new data to the video card, and since most is already loaded when you have a card with a large onboard ram size, such as 512MB on the video card, then the ram is required to tranfer data to the video card even less. The only thing left is the CPU to GPU relations, which is actually becoming less and less as new technologies are offloading more data directly to the GPU and cutting out the CPU middleman, meaning less bandwidth is required since the CPU doesn't have to send as much info to the GPU as it previously did.

As long as the video card has enough onboard space, and can store the data it needs ahead of time, then the PCIe's bandwidth is becoming less and less important. technically, a high end video card such as even the 9800GX2 could run fantastic on even PCI-e 1x as long as they use the 2GB onboard the card to good use and load everything to the card ahead of time.

 

Look at it this way. If your going to drop the green on a 8800, might as well go 2.0 than 1.0?

 

there wont be any difference, at this point. if you were to pay more for the PCIe 2.0 version of the 8800, then you've just fallen victim of a marketing scheme. Since PCIe cards right now are still nowhere near close to using up existing PCIe 16x 1.1 slot's bandwidth, then there should be zero difference.