But both tests of F1 2011 are ran with ultra settings, 1920x1080 and 8xAA.
Tom´s dozer scores 60fps, Hardware Heaven´s dozer gets 79fps.
I haven´t followed that HH site at all, should they be considered unbiased and reliable?
Does F1 2011 have a benchmark mode that both sites are using? Otherwise they could be testing different parts of the game. It's not hard to imagine one part being GPU limited at some settings and another part not being GPU limited.
What makes HH's review poor compared to THG's is that it only tests the game at one setting and it only compares the FX-8150 with the i7-2600K. If they threw in lower settings and slower/fewer core processors it'd be clearer if it's GPU limited. As it is, all you see is two CPUs getting close to the same scores over and over again, and when AMD wins by a small difference HH shouts victory. I've speculated that BD may have the advantage in some otherwise GPU limited situations, possibly due to superior uncore or GPU drivers, letting it have these slight advantages, but not when they call on actual increased CPU load.
HH leads with a preface that they only test at super high settings because those are "realistic." What isn't realistic is that everyone interested in this CPU will also be interested in buying a GPU that costs even more than it does, like the one HH used for this test. But it's still not enough to keep it from being GPU limited when they seem to go for the highest possible settings they can. If this is really all they think is realistic then they'd be better off throwing significantly slower CPUs into the mix to show that they can also reach those GPU limited points.
They also say things like this:
"At first glance the performance of our FX-8150 in video playback tasks may seem a little disappointing. That isn't the case though because what we are seeing here is the AMD CPU reducing its power consumption by moving down to a lower clockspeed than the Intel CPU which makes the performance figures look worse."
Nowhere does the review demonstrate any evidence of this being the case, nor does it even state what frequency it goes to. If this is really the case then they should have turned clock throttling off or not bothered with these tests in the first place.
Then there's the part where an overclock is shown, but despite being in the same section as power consumption there's no mention of what the 5GHz clock consumes, any indication of performance, or even any indication of stability...
Seems to me that HH should focus more on providing a thorough and balanced review instead of making it look as flashy as possible. Not saying some of the other reviews aren't weak or biased, of course. And I don't think HH is a biased review site, I just think this particular review is bad. If they have any overall negative tendencies it's that their scores are all too high.
EDIT: HH says this in a forum post:
"As for F1, no of course it isn't the inbuilt benchmark."
No further elaboration is given as to why the built-in benchmark isn't used. We don't know what THG did, but at the very least you can conclude that it wasn't the same benchmark.
still ARM cores will provide some competition eventually.
