I don't know this benchmark seems about useless. all you have to do is leave dx11 on and you can score ridiculous numbers just because you dump resolution.
case in point below. Granted I have a decent setup here, but even so I can make my number swing wildly by thousands, that just doesn't seem right.
I should also note that even in this run the min fps is still being generated by their transitions and not by the actual test which is messed up as well.
and the funniest thing is it's gone from listing my gtx580 as a intel integrated to listing it as a unknown 256mb card. It's obviously wrong on both counts.
I opened up the case and looked just to be sure