Fair enough.....cant beat them on processing power then beat them on power usage!!

So this means that for instance the X2 3800+ 32w will perform just as good as a normal X2 3800+, and overclock slightly below or just as good as one? Also when OCed, will use the same amount of power as an equally clocked and volted Non efficient X2 3800+?

Good article.

What is the exact name of program on last picture (top border)? I didn't managed to find anything with "S&m" keyword (except sado-maso :-D )

is it possible to emulate a EE AMD64 undervolting a standard one?

We should keep in mind that K8 architecture is still produced with 90 nm SOI, and Core architecture with 65 nm technology...

In the las paragraph of page 3 it says Athlon 64 X2 4600 and have 512MB of L2 instead of 2x512KB

Now that's a thorough test! Thank you, X-bit labs!

But the results make me more than confused. None of the system consumes less than 165W on idle. Of these, 30W heat the Radeon (as measured recently by X-bit). Roughly 135W remains for other components and I don't see any extremely power hungry among them.

The US Environmental Protection Agency expects to see many desktop computers with 50W consumption by 2007. Well, I really don't understand what technology will bring 135W down to 50W in a year? (Except hiding a notebook in a desktop case?)

I have a suggestion for the next test. Try to also measure power draw at the motherboard power conector, which powers only the motherboard and RAM (hopefully, given that the CPU and GPU have separate connectors). This could give a more complete picture about where all these big watts sink.

Is it possible to undervolt either AMD or Intel while keeping CnQ/EIST enabled?

You have given the power consumption results, but what about the performance. The EE processors from AMD consume less power, but do they deliver the same performance of the Intel processor that you have compared them with? Even with these EE processors, I doubt if AMD would be able to better Intel on the Power/Performance to price ratio. It would be great if you can also give the ratio also to help us understand better as to where the individual components stand.

Why did'nt they try 1.6V core for overclocking. My 3000+ is flat on overclock between 1.4V & 1.55V it wont increase more till I hit 1.6V. It will then hit almost 1.7gig. Then @1.65 it goes over 1.7 all on oem heatsink!

Sorry I ment 2.7gig on last comment!

"What you see on the chart is a complete failure of the K8 architecture."

It's not the architecture !!!
It's 90nm AMD X2 vs 65nm Intel C2D !!!
It's integrated memory controler consumption vs not adding Intel northbridge consumption !!!

Come on guys, don't be so biased...
How can you say it's a complete architecture failure ???
AMD will have 65nm chips out in 3 months...

Who cares how good these Athlon EE/SFF perform? They are late to market anyway, with a late May paper release.

Why get EE X2 right now, when the 65nm X2 will be at the same TDP (65W) in 4 months.

And the SFF versions are not even on market.

Makes we wonder why they compare to a E6300
which has only 2MB cache and low clockspeed
if they would compare to a E6600 you would see different numbers
missed opportunity
article = POOR

The article does seem a little biased towards the Intel chip. For example, in the Idle power consumption tests, I don't see the point in disabling CnQ on AMD processors while still letting the Intel chip use Enhanced Halt State.

The writing style of the article is also skewed in Intel's favor. In results where AMD processors beat Intel, the author points out an excuse for Intel. While when Intel beats AMD in other tests, the author says it's a "complete failure of the K8 architecture" or complements Intel in a subjective way.




Also, I don't this makes much sense:

"Moreover, if you recall the well-known linear dependence between the maximum CPU frequency and Vcore, everything will be absolutely clear right away. It is exactly the different Vcore that determines the maximum frequencies. Therefore, you shouldn’t believe that since Energy Efficient processors have lower nominal voltage than the “standard” ones, they will overclock better thanks to bigger range for Vcore increase. In reality if you raise the Vcore of the Energy Efficient processor to the level of the nominal Vcore of a regular one, you will simply be able to catch up with the frequency of the regular processor, but not overcome it."

If the relationship were truly linear, then you WOULD overcome the frequency of a regular processor.

It shows me that AMD's design is better than Intel's. The basic architecture of the Athlon 64 is several years old. Even so, AMD's *90nm* 3800+ X2 beats Intel's brand new *65nm* design. It should make for an interesting comparison when AMD has 65nm parts available.

WHERE CAN I BUY THE THEM AT US???

How can a AMD-CPU need more than the given TDP? Wasn't that supposed to be the maximum power the AMD-CPU can take at a worst case szenario?

To achieve approximately equal power consumption on 90nm tech is hardly the end of the K8, but rather a fantastic achievement. As a matter of fact, a double configuration of X2 3800+should give a rating of 7600+ or somewhat less and blow the core2 duos into the weeds, when fed 4-threaded software. This should hold the fort until AMD's next technology comes out.