• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Tom's Tests PCIe Scaling

klah

Diamond Member
http://graphics.tomshardware.c...ic/20041122/index.html

The question of how much bandwidth future graphic interfaces will actually need remains. In order to answer this question we played around for a little while and tested all the possible link speeds of PCI Express: x1, x2, x4, x8 and x16. To find out which of the current graphic technologies depend most on available bandwidth, we tested ATI's Radeon X800XT and NVIDIA's GeForce 6800 GT.
...
The applications we used (e.g. Doom 3 or Unreal Tournament 2003) require a minimum bandwidth of about x4 PCI Express. Above that setting, it was almost impossible to feel the performance difference. This is especially surprising considering that the rather challenging 3DMark 05, which can be described as a technical trend-setter, is not exactly conservative in terms of graphics features.

The result is even more interesting in light of the anticipated nForce4 SLI chipset, which will be available by Christmas time. This beauty either operates one PCIe graphics card in x16 mode or two similar ones (same chip and memory) at x8 PCIe each. Even though it isn't possible to double the 3D performance by using two graphic cards, there is no doubt that PCI Express will not be a bottleneck in this arrangement, not even if it is set up to operate at a slower speed such as x8 PCIe.

The difference in gaming between x4 and x16 was 0-3%. Let's just pray they don't force us to upgrade to x32 cards while they are barely surpassing the capabilites an x4 lane offers.

 
Back
Top