Yea I'm thinking it is something driver related as I've tried a couple different drivers. I might try driver sweeper or something and a fresh install of the drivers.
You won't see any real world difference with one or two cards, but quite honestly PCI-E 2.0 is why I upgraded from Z68 to X79. I had 3-way 670's and I was seeing very erratic GPU usages and most of the time usages were under 70%. I was not happy as 3 GPU"s running at 66% equals one completely wasted $400 GPU. And my results now with SLI show basically just that. My performance is not that much less at all than it was with 3-way SLI.
After reading around on the internet and finding Callsignvega's PCI-E 2.0 vs. PCI-E 3.0 thread
found here I decided that that's exactly what my problem was with my Maximus IV Extreme. My symptoms acted exactly like a CPU bottleneck just as Vega describes in his thread even though my CPU usage would drop from 45% to 22% when the GPU usages would go way down.
The interesting part was that while running PCI-E 2.0, the GPU utilization dropped way down as would be typically seen if you are "CPU" limited.
GPU utilization dropped down into the ~50% range due to PCI-E 2.0 choking them to death. As soon as I enabled PCI-E 3.0, the GPU utilization skyrocketed to 95+% on all cores.
That is somewhere in Vega's OP in that thread and pretty much describes exactly what I was experiencing. While his usages were in the 50's due to 4-way SLI I mostly saw usages in the 60's and 70's with 3-way. Still very frustrating to look up and see 66% or less usage across the board and know that my 3rd card is literally doing nothing for me.
So short answer no you won't see real world difference with one or two cards and single screen resolutions, but once you start pushing 3 and 4 cards and triple monitor resolutions you will be bottle necked by PCI-E 2.0. It's not necessarily a huge deal for me right now as I had to trade away the third card for the processor, but if I ever went 3-way again I would definitely want PCI-E 3.0 to be running properly.