It's related to Intel's FSB vs AMD's HT.
Don't worry, future Intel chipsets will take care of the problem :)
Printable View
http://images.anandtech.com/graphs/r...3626/17190.png
WTF happened here?! CF actually slower, yet X2 scales like a dream? I guess drivers were optimized for X2 only.
or he has reused scores from previous testing with older drv?
I just published my ATI Radeon HD 4870 X2 preview on Finnish hardware and overclocking site Muropaketti.com:
http://plaza.fi/s/f/editor/images/X-...1333482181.jpg
http://plaza.fi/muropaketti/artikkel...d-4870-x2-r700
Small Finnish to English dictionary:
Minimi = Minimun
Keskimääräinen = Avarage
Maksimi = Maximum
Article includes my tests with microstuttering issue which I'll translate to English later (hopefully today) with graphs and raw measured numbers.
The 4870x2 performs worse in general right now then 4870 in CF. Oh noes. Who called that one. I will admit that They mainly perform the same now, because of a general better design. Where do you guys see 70% Age of Conan? Okay? and when did that become a standard to BENCHMARK. I noticed also, Why isn't there any 3dmark06 benchs in the previews.
When I look at benchmarks, I expect them to be like HL2:EP2, TF2, 3dMark06, 3dMark Vantage, Quake Wars, Crysis, Call Of Duty 4, FEAR, NFS:Pro Street, UT3, stuff like that, the only thing I seen test was Age of Conan in all of the reviews mainly, Crysis, and GRID. Yes GRID is pretty. But it seems like a one-sided benchmark on it.
http://plaza.fi/muropaketti/artikkel...4870-x2-r700,1Quote:
The 4870x2 performs worse in general right now then 4870 in CF. Oh noes. Who called that one. I will admit that They mainly perform the same now, because of a general better design. Where do you guys see 70% Age of Conan? Okay? and when did that become a standard to BENCHMARK. I noticed also, Why isn't there any 3dmark06 benchs in the previews.
When I look at benchmarks, I expect them to be like HL2:EP2, TF2, 3dMark06, 3dMark Vantage, Quake Wars, Crysis, Call Of Duty 4, FEAR, NFS:Pro Street, UT3, stuff like that, the only thing I seen test was Age of Conan in all of the reviews mainly, Crysis, and GRID. Yes GRID is pretty. But it seems like a one-sided benchmark on it.
3DMark benchmarks.
And if you actually seen more than one review you would know that reviewers could only pick 4 games which they can bench. Which ones is up to them.
http://firingsquad.com/hardware/amd_...view/page3.asp
fs does, tought not up to 1920x1200 but 1600x1200 with is quite enough to.
I really think something went wrong with that review.
Hmm mybe somone with a phenom rig and a HD4870 can do some testing and a guy with a penryn rig. Should be possible to pull this off.
Anyways enough offtopic. ;)
and if you actually looked at my post, there was many preview. Don't tell me that, I've seen Reviewers have 6-8 games before.
Also, I forgot Bioshock and Devil May Cry 4 are good ones to benchmark.
Those 3Dmark06 don't seem very actuate. But the article is hard to read since I don't know anything other then English.
I'm saying there is a lot better games that could have been benchmarked if they was really limited to 4.
I don't see any problem with Benching Conan. I don't play it myself but a lot of people do and it is quite new. Rather Conan than Fear TBH.
The card performs pretty much as expected for me. If it's cheaper than 2 4870s, then it's a decent card. Bit expensive for my liking but then again so is the 280.
If I had $600 to spend, then the 4870x2 would be the choice I would take followed by the 280.
So what it's popular, The One I stated are still used as benchmarks.
and When did Devil May Cry 4 become old :confused:
I wouldn't mind them using Age of conan if the DX10 feature worked, That would add something to the table. But it's not DX10 Enabled yet. What happened to Assassin Creed. :rofl:
Things like Age of Conan and Crysis are tested because they are among the most demanding titles out now ( and likely for some time ) I do agree MMOs are very hard to bench as the worlds are persistent and you won't ever get the same exactly run twice ( for this reason I wish all MMOs had a time demo for testing purposes ) and Crysis doesn't seem to like crossfire but these still are demanding games, both in memory usage and shader requirements.
@Warboy - As far as DX10 and Conan we won't see this until late fall ( its going to be demoed next month but won't be finished and realeasble for another couple months according to the game director ) Regardless (even without DX10 ) Conan is still quite demanding and not a single reviewer has done a worse case scenario test with it yet, to be fair that would be very hard to test ( HardOCP runs through Old Tarantia, which can be very demanding, but a large raid would be even more demanding, to the point of bring a 4870x2 in the 10-15fps ranged maxed out I'd bet ) The problem with MMOs is the consistency of benchmarking for the reasons I mentioned above. Based on player load you can get some pretty substantial variations in testing but I guess a run through a crowed city during peak times will give a decent enough baseline.
Warboy, just enjoy your GT280. Does it really matter for you if the R700 is 70% faster than the GT280 or not? Cause from the way I read your posts, you are anti AMD anyways. And the GT280 is fast enough even though it is slower than the R700
To be honest so far i find HardOCP preview to be most informative. Then again i stopped reading after Hexus, just waiting on Sampsa's english translation.
Eh. To me, the 4870 X2 seems to be the winner all around.