Guys, it seems like your test is screwed up. Just check any other 4850 Crossfire tests, they all generally show improvements in all the tests where you state that "crossfire doesn't function here". Especially in Call of Duty 4 where the card scales extremely well resulting in almost 100% scaling improvement. Check Anandtech and Guru3d's tests

I have just counted 3 games that your
results are so wrong : bioshock,Call of duty 4,
S.t.a.l.k.e.r , i havent bother to look at the other.
This is horibble, you have to bring the review down,
and remake it right,this is for laughs

To enable CrossFire you must go into CCC, select the CrossFire tab and enable the feature. This is a MANUAL procedure - there is no warning at all that you have a CrossFire capable board or that you have to go into CCC to enable CrossFire. Your review don't reflect the reality, in call of duty 4 you must obtain almost a +98% performance, not -5 % like in your review.

Um wtf . . you telling us that CF doesn't scale on COD4, but what about this review then from Guru 3D ?
>>

What . . how come these scaled in COD4 then on Guru3D ???

[url]http://www.guru3d.com/article/radeon-hd-4850-and--4870 -crossfirex-performance/5[/url]

The problem is the motherboard, when in crossfire mode its 8x + 8x. Which is bottlenecking the cards. Xbit should have use a motherboard with 16x + 16x crossfire (one that uses intel x38 or x48 chipset). If you check the other review sites they use X38 or X48 chipset motherboard. The DFI LANParty UT ICFX3200-T2R/G, that xbit used can only do 8x + 8x in crossfire. They should redo the review with a better motherboard

"Relatively high power consumption" = at least a 750watt PSU, which will be a required upgrade for most people. Add that to your cost considerations...

xbitlabs is only good for CPU reviews.

You guys need a new testbed that's free of any bugs and a guy who knows what he's doing, not some noob with a computer degree. Also check your numbers before publishing them since people make their purchase decisions on these numbers.

the Print button should open an new box not Use the page that is open (how every one els is doing it)
happy that print button is there now do not need to use the next button any more

i have to agree the results on this side to do follow what other sites CF results have shown (8x PCI-e 2.0/1.0 does not limit the results but motherboard problems can)

i got an intresting problem with with 2 pcs when i turn on SLI (8800gts 320mb) i get 300 points less in the SM2 and SM3 tests in 3dmark06 (does this as well with 8500gt cards as well)

de Benchmarks are not correct.. whats up with you reviewers at Xibit, maybe you guys just need to stick with CPU reviews instead of Gpu

yeah get a new mobo for the CFX....

and abt the accessories u reviewed
i found a retail sapphire hd 3850 card over here and ur hd 3850 review said dat it included a version of 3dmark06 and black box game but wen i checked the retail package it didnt have 3dmark06 or the blackbox game... so u better ask sapphire if the package they send u will is the final retail version or not...
i think the same applies to the hd4850 retail package...

After 5 days of critics, 2 or 3 new reviews, this is still the first review in your page. Are you Nvidia boys ? as Mr. BonBon seys, people make their purchase decisions on these numbers.

While you seem to have trouble getting Xfire working on some games at least you included the min fps (which guru didn't). I'll take a lower average but higher min fps over a higher average fps any day.

I would like to see a review of two individual 4870 cards, where the memory issue goes away and we can see what the cards actualy do in comparison to the rival. Also add in Supreme Commander to the list of reviews as most other review sites include it for comparison.
Something is definatley wrong with the COD4 test. It should have been way better that your review. You need to do a sanity check before printing that crap, obviously something was wrong with your test result.