Nehalem takes less than a volt

[PP Mguire]

Check it out.
http://www.tomshardware.com/news/intel-x58-nehalem,5829.html

[ahsalloum]

Wow, that's impressive.

[joz]

one volt, less then a volt? what is the LV and ULV versions of this beast gona be like?

[PP Mguire]

I think if they stopped trying to decrease volts whilst increasins performance they would get ALOT more performance.

[joz]

but for servers, performance/watt is the most imporatant thing. and the lower the volt, the lower the wat.

[bobletman]

Wow that is amazing. HOLY CRAP nehalem is going to touch AMD.

[Goliath182]

I think if they stopped trying to decrease volts whilst increasins performance they would get ALOT more performance.
Low energy usage is the new craze lol. As long as it performs better than Conroe im happy.

[Acidtears33]

I think if they stopped trying to decrease volts whilst increasins performance they would get ALOT more performance.

Not all the time. As transistors get smaller and smaller their capacity for voltage diminshes. This mean that while they can operate at the same speed as larger transistors but use less energy, it also means that the same voltage applied to the larger transistors (overclocking type) can damage them.

ie. A 65nm C2D can safely take up to 1.55 volts without immediate fear of damage. Shortened lifespan not withstanding.

A 45nm C2D can safely take up to 1.45 volts.

So the reduction in voltage is sometimes a necessity. Though I would agree that from one generation to the next they would be able to keep the same amount of voltage and operate safely.

Could you image if they pumped the same amount of volts through these things as a PIII?!?!?! :jawdrop:

[Kougar]

Well, early benchmarks and power tests show that Nehalem runs pretty cool, and within the same power envelopes that current 45nm Quads fit within.

If you think about it... I expect Intel was only able to maintain the same power envelopes by shooting for a lower voltage, hence why these chips are being seen running full speed using 1 volt (give or take vdroop).

It should mean there is plenty of overclocking headroom with voltages of just 1.20v though... :grin:


Could you image if they pumped the same amount of volts through these things as a PIII?!?!?! :jawdrop:

Someone did that to a 45nm QX9650, fed it 1.7-1.8v and got around 5.5Ghz give or take a few hundred. Since it wasn't subzero cooling I'm amazed the chip still works.

[Goliath182]

A P4 631 (3 gigs) has been over clocked to 8.2 gigs, it was pumped with a massive 2 volts. Imagine if you could do that with Nehalem. :ahhhhh:

[Xantor]

that's pretty awesome. power efficiency should be top notch as well!

[Kougar]

Voltage isn't nearly as harmful if the CPU is kept below 0c. Remove the heat, and voltage by itself is much less harmful since the additional heat it creates is wicked away instantly.

I was almost able to grab a dry ice pot awhile back from one place, could've gotten temps somewhere between -40 to -70c with it on a Q6600... I would likely have tried 2v on my E6300 if the temps were -70c or colder. :)

When you hear about someone using insane voltages... it's not insane unless they are using air/water and not subzero cooling of some sort or another!

[PP Mguire]

Well we all know that. My point is if they used more volts i bet they could ramp the clocks up and really hit home with the performance. Idk if im the only one, but im really not liking the return of Netburst and HT.

[peti1212]

I don't get it, what's so wrong about the HT? I heard from others too that they are not liking the idea.

[PP Mguire]

What would you rather have a wanabe 8core or a real 8 core?

[peti1212]

Oh, I see what you mean. It is still better than a quad though. Just like the P4 it had HT, and you could run two folding @ home clients on it. One would for example put out a WU in 2 days while the other HT thread would put out 1 in 3 days. So I would say it will work like a 6-7 core processor.

[PP Mguire]

Eh dosent work that way. The old P4s with Netburst sucked ;) I have 2 sitting in my closet and boy does my AMDs run circles around them.

[peti1212]

I would still suspect it being much faster than the current processors.

[Goliath182]

Eh dosent work that way. The old P4s with Netburst sucked ;) I have 2 sitting in my closet and boy does my AMDs run circles around them.
Thats due to architecture not Netburst. And now a days any new Intel would smoke a AMD. :lol:

[PP Mguire]

Thats because AMD is lacking behind like sore losers.

[Goliath182]

Thats because AMD is lacking behind like sore losers.
Because they didnt have money after buying ATI.

[PP Mguire]

They didnt have money before Athlon either ;)

[borschtBomber[SS]]

lol, I don't think AMD has ever had money...but lets hope they come back strong and keep Intel in check ;)

[Goliath182]

They didnt have money before Athlon either ;)
True but there still calling their quads Phenoms. :(

[3991vhtes]

This is great for intel, because I would rather have a low-power CPU, and a more energy efficent computer.

Now, I wonder when AMD will catch up? I don't like their technology (the quad core phenom's were a joke), or their fanboys. Socket A is the ONLY AMD you'll see me use.

[PP Mguire]

Socket 939 is the best AMD socket to date.