No, they just do not measure with FCAT. Meaning the spikes you see are internally measured in the game engine, not what you see with your own eyes. E.g. they measure the rendering engine and not display output. This is the sole reason why FCAT was introduced as it plots what you actually see. http://www.guru3d.com/index.php?ct=articles&action=file&id=15352 The article has been updated with extra UHD results and fcat you guys. Indeed a tiny hint more frame pacing latency in-between frames, nothing dramatic though.

I cant find mantle benchmarks anywhere. Does anyone know why? Sorry for my english

Mantle benchmarks

Hilbert used 15.3 Beta compared to others 14.12 (omega) There should be any improvement, no ?

It will be interesting to see what the benchmarks show when Windows 10 is released. I believe both Nvidia and AMD are working on their WDDM 2.0 drivers more so than making their Windows 8.1/7 drivers better. As I said in another thread, realistically this is fair enough, because doing heavy work on Windows 8.1/7 drivers now is kind of pointless and Microsoft would prefer they don't anyway. If you want optimised drivers, just take Microsoft up on the free upgrade offer. Why should Microsoft, AMD, and Nvidia spend time and resources continually developing for stubborn people that can't accept change 😉. The driver support doesn't affect business running Windows 7, Server 2008 R2 etc, since basic driver support can still be provided and games are a non-issue for them. Windows 10 won't show a true representation just yet, as they are still in their infancy, and Windows 10 isn't feature finalised or full feature optimised as yet.