I'm not buying at Microcenter, because I live in Europe

.
But all of those tests are at 1680x1050, wouldn't there be a difference at 1080p?
Uh... yes and no, but for all intents and purposes it's a very resounding no.
Gaming benchmarks are done on 2 scales: low end and extremely high end. The reason for that is due to the reviewer correctly eliminating the GPU as a factor in the tests (as much as possible) and that's either done when you tune your resolution and settings low or it's done at the very high end where (usually) multiple GPUs are tested at very high settings and resolutions such that the processor becomes the bottleneck. On both accounts the new Bulldozer chips lose to their Intel counterparts and the 2 and 3 module chips lose out to 4 and 6 core Denebs/Thubans.
Sorry, I meant 4.0ghz on 4 cores on a 960T seems to be quite common. I'd guess the 3.6-3.7 range on 5/6 cores would be the expected range.
At 1080p you're likely GPU limited, thus nearly all the mid/high end desktop processors will perform the same and therefore makes a worthless benchmark. There is always a slight difference, but that difference is usually magnified when lowering or increasing settings/resolution. Some game devs tend to favor one CPU manufacturer (or architecture) over another, and a prime example would be Blizzard's titles. Although they're threaded, albeit very poorly, they tend to favor higher IPC and as a result you'll see the 8150 losing to a core i3 or SB Pentium.
In general, if you want gaming performance from a CPU you'll go Intel. If you're buying AMD then skip over Bulldozer and get yourself a Thuban or Deneb for cheap. The stock is running dry on the 6 core Thubans, but the 960T is a fantastic chip for cheap.