My 480SLI setup pretty much matches a fellow forumites
quadfire setup in Dirt2. I guess we are both evil biased reviewers on the payroll of the GPU manufacturers.
http://forums.overclockers.co.uk/showthread.php?p=17713399#post17713399 Now were do I collect my green cheque?
Out of interest, which sites can we trust these days? It seems the list of green tainted sites is a moving target this week
I don't hold much truck with any review site these days... very few seem to use the right drivers for each make, rarely any attempt to match driver settings i.e. "Good Quality" settings can be different between AMD and nVidia in regards to what they do and lastly very few use frametime analysis which would tell a far more complete story than min/max/avg which are easily skewed.
AS Rroff said, my issue with review sites isn't in general green or red bias, its just, being a bit crap, of which most are. Most are painfully lazy about their methodology, even worse some just draw ludicrous conclusions that have no basis in reality. realistically a lot of these sites simply don't have the hardware, while a decade ago AMD/Nvidia would send out cards left right and centre and anyone who starts their own review site could get freehardware, it changed dramatically. Lots of places get a card for 2-3 days and then have to send them off to another review site or back to whoever they got it from.
So some sites have old results because they don't have the hardware to get new results for every new release with new drivers, the difference in quality in those sites(of which there are many these days) are those who tell the truth, who tell you what drivers are being used, will say they don't have card X anymore so on newer drivers you can expect X performance from AMD/Nvidia cards compared to 6 months before. Very few do that though, most lie through the teeth and you can see, especially on Hardware Canucks, that they claim to use new drivers but the results are clearly ripped out of 6 month old reviews.
Which incidentally is why the 460gtx reviews look so good, they use old 480GTX numbers aswell so the 460gtx looks far faster compared to AMD cards AND the 480/470gtx than it really is. THe reality is, 90% of review sites out there, are completely rubbish. Another 5% will happily lie and use dodgey results to favour not really one company over another, just whoever will throw them the most advertising cash for reviewing their cards. 98% of sites make money through ad sales direct from the guys they are reviewing cards for, and if you don't think theres some, incentive for making a card stand out or recommending it /scoring it better than it is, you're mad.
Theres also sites out there who I've heard(from very good sources, keeping in mind I used to work for a review site and have met loads of industry/pr people) will simply refuse to review things if companies don't buy up X amount of advertising space. They weren't biased, they wouldn't have reviewed it better or worse, they were just greedy gits.
Your sli might match quadfire, but how much faster would quad sli be. Quadfire/quad sli scales like crap. So your sli matches quadfire, but how much ahead of xfire is it. Double the speed or only 10% behind. The 480gtx is between 10-20% faster in general, why wouldn't SLi be ahead?
Techreport as I've highlighted, removed EVERY SINGLE AMD dominating benchmark, except 2 or so, one of which was Call of Juarez, its marginally ahead of a 5870, but the same benchmarks, on a 6 month old review, with the 480gtx getting the same results within 1%, and the 5970 is something like 55% faster, or maybe 70% faster(i forget which is which offhand).
They still include Call of Juarez 2, and the 5970 is included in the review, but for that benchmark and one other the 5970 ONLY was removed. So in a review they removed say 2 results while including 5970 results in another 15 or so benchmarks, and just happen to remove the 2 with the biggest lead for the 5970?
Sorry but you really can't class that as anything but insanely dodgey.
Its akin to say Anandtech reviewed the 6870, but left out 3 "standard" benchmarks for the past two years that we knew showed the 480gtx 50% faster than a 5870, and then left out the 480gtx result in a couple more.
Its laughable, infact I've never seen another review site compare say 10 different cards in a review, and for 2 of the benchmarks, leave out results for one card AND THEN draw conclusions and show the speed difference leaving out the results making the difference quite deliberately look smaller.
