I remember reading the tom's article on bandwidth of the 8xxx nvidia cards on different number of lanes, and remember that only the 8800GTX actually lost a bit of performance using a 4x PCI-E 1.0 slot.
Now with the next gen cards, the cards like the 4870X2 and the GTX280 that are designed for PCI-E 2.0 (32x), How do they perform on a PCI-E 1.0 (16x) slot? (non sli of course)
Can tom's do a review on this for us? I'm sure i'm not the only one with a pre PCI-E 2.0 board.
if 4x was too slow for the last gen, I'd say x8 should be the cut off. With all the MB supporting 2.0, 8X runs as fast on these new cards as 16X did on the GTX. We should have some breathing room for a while. (at least until 3.0 in 2010)