I think that alot of you guys should go back and read the article again. It is very clear that the DFI Ultra board that Wesley Fink tested was modded to be an exact duplicate of the DFI SLI marketed board. The Ultra was tested at x8 x8 and not just the x16 x2 that is stated here several times. And make no mistake the performance is identical. Here is a quote from the article stating as such "We first confirmed that test results were the same on the LANParty UT modified to SLI and the LANParty nF4 SLI, which is a native SLI chipset board. There was no difference in performance after the SLI modification to the Ultra chipset, so results are reported as SLI and relevant to either SLI or Ultra modified to SLI" The x16 x2 results were included just to see what the penalty would be for that configuration. Of course, probably no other manufacturer will release an Ultra board with two x16 PCIe slots and the x16 x2 results would represent the performance of those other Ultra boards with the mod. Hell, DFI probably won't release this one if Nvidia has anything to say about it. I'm curious to see how this plays out. Any bets on Nvidia putting the sqeeze on manufacturers that try to use an "Ultra" chipset on a SLI pcb with two x16 slots?