Under FSX, there's almost a 20-30 FPS difference (with BP=0) between PCIe 1.0 and 3.0.

 

The games in the test I linked to leaned very heavily on the graphics card and showed little difference, especially between PCIe 2 and 3. In FSX, where the CPU does most of the work, you'd expect the difference to be even less, not much more. Also, the test used exactly the same motherboard which, unusually, could switch the x16 slot containing the graphics card between the different bus standards, so eliminating any effect from using a different chip set or BIOS. In the thread that you linked to it's not clear whether the same motherboard was used for each test - unlikely as the ASRock Extreme4 was mentioned and that would only appear to be able to run the PCIe 3.0 slots at x16. If it was done on different motherboards, the results would not be a true comparrison. Maybe I'm missing something here?

...has anyone heard when the next nVidia graphics cards will be coming to market or are the 980TI and Titan X the top cards for a while...?

In the thread that you linked to it's not clear whether the same motherboard was used for each test - unlikely as the ASRock Extreme4 was mentioned and that would only appear to be able to run the PCIe 3.0 slots at x16. If it was done on different motherboards, the results would not be a true comparrison. Maybe I'm missing something here?

 

There's a setting in the BIOS of the extreme chips that allows you to change the PCIe bus link speed from 1.0 up to 3.0.

 

The regular chips don't have such a setting.

 

There is also some tweaks from Nvidia that do this when PCIe 3.0 first came out.