G0 stepping - only 48 deg C when overclocked ?

[NoobyDoo]

HotHW Link

<div class="FTQUOTE"><begin quote>
This time around, however, our results were nothing short of spectacular. By bumping the processor's core voltage to 1.45v and increasing the front side bus speed via our Asus P5K Deluxe motherboard's BIOS, we were able to take the Core 2 Duo E6750 up from its default clock speed of 2.6GHz to an impressive 3.92GHz (multiplier=8x / Front Side Bus Frequency=490MHz. This was done using the stock Intel CPU cooler on an open-air test bench. At that speed, the CPU completed a Cinebench rendering pass in just 18 seconds and it blew past the Core 2 Extreme X6800 in SANDRA's Processor Arithmetic benchmark.

We should also note that throughout all of our overclocking experiments, the CPU barely hit 48°C and it seemed to heat up and cool down very quickly. We've only spent a limited amount of time with this chip and haven't experimented with multiple motherboards just yet, so we're not certain these temperature readings are accurate, but this early data suggests that this new revision of Conroe runs quite cool and has excellent overclocking potential.
</end quote></div>

See also :
http://www.anandtech.com/cpuch...el/showdoc.aspx?i=3012

http://www.xbitlabs.com/articl...uo-e6850_11.html#sect0

http://www.techreport.com/revi...uo-e6750/index.x?pg=14

Edit : Looks like the HSF manufacturers will join AMD on the endangered species list ;-).

 

[MarcVenice]

heh people could be hitting 4ghz, on air I don't understand why Intel doesn't release cpu's that run at 3ghz and higher, straight out of the fab though ? It will increase the gap with AMD even further, especially if they don't ask ridiculous prices for it. I mean, if someone could buy a 3gh intel c2d for the price of a x2 6000+, the choice would be fairly easy.

 

[F1N3ST]

Well, Penryn (45nm C2d) is supposed to have clock speeds over 3Ghz.
Also I see a pattern.

180nm, 130nm, 90nm (half 180), 65nm (half 130nm), 45nm Penryn (Half 90NM), 32nm Nehalem (Half 65nm), 22nm is after that by my guess but who knows

 

[AmpedSilence]

Originally posted by: MarcVenice
heh people could be hitting 4ghz, on air I don't understand why Intel doesn't release cpu's that run at 3ghz and higher, straight out of the fab though ? It will increase the gap with AMD even further, especially if they don't ask ridiculous prices for it. I mean, if someone could buy a 3gh intel c2d for the price of a x2 6000+, the choice would be fairly easy.

Why should they? AMD didn't release anything that would increase the gap by too much when they were in the lead? Intel has no competition, once AMD can bring some competition, they can just bump up the speed and continue to hold the performance crown. AMD did the same thing...

 

[RaynorWolfcastle]

Originally posted by: F1N3ST
Well, Penryn (45nm C2d) is supposed to have clock speeds over 3Ghz.
Also I see a pattern.

180nm, 130nm, 90nm (half 180), 65nm (half 130nm), 45nm Penryn (Half 90NM), 32nm Nehalem (Half 65nm), 22nm is after that by my guess but who knows

Actually, the pattern is really straight forward, each node has feature sizes approx sqrt(2) times smaller than the previous node. So that if you could make a "dumb shrink" of your circuit from one node to the next, your circuit area would be cut in half. That is if you have a circuit that is 1 cm x 1cm at 65 nm, it will be (roughly) 0.7 cm x 0.7 cm, or 0.5 cm^2 at 45 nm.

 

[coldpower27]

Originally posted by: MarcVenice
heh people could be hitting 4ghz, on air I don't understand why Intel doesn't release cpu's that run at 3ghz and higher, straight out of the fab though ? It will increase the gap with AMD even further, especially if they don't ask ridiculous prices for it. I mean, if someone could buy a 3gh intel c2d for the price of a x2 6000+, the choice would be fairly easy.

Multiple reasons, it's easier on the yields for one if they keep the clockspeed constant over such a long period of time.

They already make plenty, of moola with their lead now, so it's safer to hold some reserve.

They are also pushing perf/watt so it's hard to increase the frequency that much farther while staying in their 65W envelope, I am sure they could do 3.6GHZ if they wanted to, it will probably just be 95W TDP or higher, which would go against Intel's perf/watt initiative.

 

[SerpentRoyal]

Originally posted by: NoobyDoo
HotHW Link

<div class="FTQUOTE"><begin quote>
This time around, however, our results were nothing short of spectacular. By bumping the processor's core voltage to 1.45v and increasing the front side bus speed via our Asus P5K Deluxe motherboard's BIOS, we were able to take the Core 2 Duo E6750 up from its default clock speed of 2.6GHz to an impressive 3.92GHz (multiplier=8x / Front Side Bus Frequency=490MHz. This was done using the stock Intel CPU cooler on an open-air test bench. At that speed, the CPU completed a Cinebench rendering pass in just 18 seconds and it blew past the Core 2 Extreme X6800 in SANDRA's Processor Arithmetic benchmark.

We should also note that throughout all of our overclocking experiments, the CPU barely hit 48°C and it seemed to heat up and cool down very quickly. We've only spent a limited amount of time with this chip and haven't experimented with multiple motherboards just yet, so we're not certain these temperature readings are accurate, but this early data suggests that this new revision of Conroe runs quite cool and has excellent overclocking potential.
</end quote></div>

See also :
http://www.anandtech.com/cpuch...el/showdoc.aspx?i=3012

http://www.xbitlabs.com/articl...uo-e6850_11.html#sect0

http://www.techreport.com/revi...uo-e6750/index.x?pg=14

Edit : Looks like the HSF manufacturers will join AMD on the endangered species list ;-).

I'm not sure about the accuracy of that 48C temperature with the stock cooler. Which temperature monitoring software? Which stress testing software? What's the ambient air temperature?