AMD Beats Intel in 36 of 57 Power Efficiency Tests

[covert24]

http://www.worlds-fastest.com/wfz991.html
Chicago, IL, August 30, 2007 -- Recent tests by Neal Nelson & Associates, an independent computer performance consulting firm, have reported that in 36 of the 57 cases tested an AMD (NYSE: AMD) Opteron based server delivered better power efficiency than a comparably configured Intel (NASDAQ: INTC) Xeon based server

 

[Lonyo]

Mainly due to Intel's choice to use FB-DIMM instead of DDR2, since that causes serious power issues when it comes to idling, and also when you start to increase the amount of RAM you have.
Anandtech did a similar test a shot while ago IIRC.

 

[Arkaign]

Yep, it's more the platform than the processor for those tests. Good stuff for AMD there, however, and it remains their sole remaining shining spot. Not enough to remain viable, though. They need to raise ASP to profitable levels, so that the next 3 quarters don't see ANOTHER 2 billion dollars in losses. AMD will go the way of 3dfx and Cyrix if this keeps up for another year.

 

[Keysplayr]

An Intel representative commented: "The report doesn't measure our latest Xeons, or quad cores. We have 2GHz quad cores in the market at 50 watts, 12.5 per core! The report ignores performance, in that you'd use less Intel servers to get the same job done, meaning less electricity is needed.

"We stand behind all of our energy efficient claims, period. For those IT managers who don't do their own in-house testing, we recommend that each look at the 100s of independently verified benchmarks and reviews that exist for the most credible assessment."

Anyhow, it's the FBDIMMS that cranks up the power usage for the platform. They really do need to get away from them. It's like Intel & RAMBUS all over again. Trying to revolutionize the memory frontier. Give it up already. ECC DDR2/3 would suffice.

 

[Griswold]

What good is in what they say afterwards when the white paper clearly stated that Intel was not interested in talking to the NN&A people about the test, before, during and after. Kinda of typical Intel stunt, if it doesnt make them shine, they ignore it and try to belittle it when its too late.

Is there a source for that "representative comment"?

 

[Nemesis 1]

Originally posted by: Griswold
What good is in what they say afterwards when the white paper clearly stated that Intel was not interested in talking to the NN&A people about the test, before, during and after. Kinda of typical Intel stunt, if it doesnt make them shine, they ignore it and try to belittle it when its too late.

Is there a source for that "representative comment"?

You don't need a link the article tells what processors were used. The good part is as Keys said. Intel doesn't need Fbdimms right now. Because its a weapon that AMD is using. Which to me is just wrong. Once FBdimms get perfected thank god for intel because without them we still be using DDR dimms. As AMD always waits for Intel to take new advances and works threw the problems than and only than does amd come on board. FBdimms may not be the best performance or energy efficient dimms. Right now. But as they become more energy efficient they will be pretty good dimms for servers and workstations. The best part is Penryn will have a server based chipset that doesn't use FBdimms. So intel will have the best of both worlds unlike amd which will be stuck on ddr2 memory till 09.

 

[Viditor]

Originally posted by: Nemesis 1

Originally posted by: Griswold
What good is in what they say afterwards when the white paper clearly stated that Intel was not interested in talking to the NN&A people about the test, before, during and after. Kinda of typical Intel stunt, if it doesnt make them shine, they ignore it and try to belittle it when its too late.

Is there a source for that "representative comment"?

You don't need a link the article tells what processors were used. The good part is as Keys said. Intel doesn't need Fbdimms right now. Because its a weapon that AMD is using. Which to me is just wrong. Once FBdimms get perfected thank god for intel because without them we still be using DDR dimms. As AMD always waits for Intel to take new advances and works threw the problems than and only than does amd come on board. FBdimms may not be the best performance or energy efficient dimms. Right now. But as they become more energy efficient they will be pretty good dimms for servers and workstations. The best part is Penryn will have a server based chipset that doesn't use FBdimms. So intel will have the best of both worlds unlike amd which will be stuck on ddr2 memory till 09.

Arrrrrrggghhh...NO!

1. Intel only needs FBDimms if they want to use large memory arrays...and I can't see them not wanting that. It's more important than the power costs!

2. FBDimms aren't new, they've been around for quite awhile (IBM has used them for years). They really aren't going to get much better from a power perspective, and remember that they use DDR inside as well.

3. Intel will be moving away from FBDs when CSI is released on Nehalam...the reason they use FBD and AMD doesn't is due to the difference in system architecture (so you could say that it's Intel that was waiting on new advances in this case). P2P (like HT and CSI) allows for far more memory slots on the system than a FSB model does, and the latency is necessarily MUCH lower (no way to engineer around that).