Nowyoudie

Yes. Of course, before it bursts into flames and pops through the motherboard on the back.

__________________
iPhone Club

Blameless

Not true.

Voltage damages ICs (inluding CPUs) through electromigration.

Electromigration is highly dependent on electrical resistance.

Less heat = less resistance = less electromigration; even if the voltage is the same.

1.3625v is 24/7 safe for 45nm chips up until about 70C.

If much colder temperatures are maintained, you can achive the same lifespan with much higher voltages.

Since the difference between good air and good water is in the ballpark of 10C max, the difference in safe volts is not huge. However, if you have something like a chilled water TEC, a cascade, or LN2, you can maintain temps that are more than 100C cooler than what you would get on air. This allows for a much higher safe voltage.

A well established rule of thumb is that for every 10C you reduce temps, you double the lifespan of an electronic component. Going from 55C max load to 45C max load at, say 1.5v, could allow a 45nm Intel to last for two weeks instead of one. Dropping from 55C to -45C would theoretically increase that one week life span to 1024 weeks. Now the rule of thumb is not going to hold up under such a huge range, but even if it's off by a factor of two or three, your chip will still last several years.

It's all about the change in absolute temp. 55C to 45C is nothing. It's actually going from 328K to 318K, or a few percent change in absolute temp. Going from 55C to -45C is going from 328K to 228K. That's pretty big.

EDIT: I should also add that reducing resistance will also allow you to clock higher on the same volts. Lowering temperatures reduces the voltage needed for a given clock and allows you to safely use higher voltages.

With good enough cooling, 4.8GHz could very well be possible, on a good chip, with stock volts.

-150 to -130C.

sLowEnd

Such high voltages required for these kind of OC's will cause damage to the chip regardless of cooling. The cooling just makes the chip last long enough to run for a bit, but it'll still be permanently damaged

__________________

DarkVader

Well that didn't last long

Here the new record by Boblemagnifique submitted 12 and half hours later



Validation

40Mhz difference

Blameless

That's true to an extent. Even a 100C+ drop in temps won't save a wolfdale from something like 2 volts.

It's just often the rapid cooling (and warming) that kills components in these sort of OC experiments.

Take a totaly stock chip, attach a LN2 pot to it, rapidly cool it to 150C for a while then let it warm back up, and there is a good chance of the chip being dead, even if you never bumped the voltage at all.

antd

6GHz is nothing compared to this 10GHz quad

http://cgi.ebay.co.uk/Desktop-Gaming...6.c0.m14.l1308

alk

ahahaha! Don't you just hate it when people do that! Clockspeed does not equal number of cores multiplied by clockspeed! hahaha! Why would anyone buy a PC off that guy!

If I wear 2 wrist watches, does time go twice as fast?!

__________________

Currently studying for an MEng in Mechanical Engineering - 3 years down, 2 to go!

Hewlett-Packard

Dare to dream sweetheart.

Will CPU's ever be able to reach this high of a clock normally? Or are we just going to keep multiplying cores forever?

__________________

||||||-------------------

$300/$1200

antd

lol

you were wrong!

He clearly states: 10.6 Ghz Computer Tower : The CPU speed shown is the GHZ when four cores are running efficiently.


I strive for that efficiency!! The sad thing is that he has good feedback on ebay, in fact 'he' is an ebay store. People fall for this =/

onlycodered

Woah. That's like twice as fast as mine.

__________________