$2000 for new system

[centvalny]

I'll believe it when I see it. Such statements violate basic microprocessor design principles.

Heres..from 2.6Ghz stock..

 

[(cf)Eclipse]

"10c lower = 100 MHz higher" is also complete and total disinformation, and whoever told you that, has no clue what they are talking about to begin with. Every chip has its own clockrate ceiling, and you can cool a chip all you like, it won't pass it.

look. this is getting rediculous. i have done lots of testing, i KNOW that colder temps = higher stable speeds. end of story

 

[DejaWiz]

It's proven that colder temps = less electrical resistence = greater circuit efficiency.

The Meissner Effect. The BCS Theory. These are associated with the practice of supercooling certain substances in hopes of allowing perpetual electron flow. Also known as superconductivity.

Learn more about this at http://superconductors.org/

Now back to the topic. An overclock of 2.6-2.8GHZ won't yield much net gain in overall real world performance over 3GHz. The only thing that will really be gained is, of course, bragging rights. But is the risk and investment worth it? To me, no not at all. I've got my 3000+ Venice clocked to 2.4GHz (nothing to really brag about). On a Gigabyte board. Now that's a heck of a risk LOL. Going from 1.8GHz to 2.4GHz is a decent gain in speed, but because the Athlon is so efficient to begin with, it's not like a 3GHz Venice would be leaps and bounds faster than one running at 2.7-2.8GHz. Here's a benchmark comparing a Venice clocked at 2.0GHz and 2.8GHz. The gain is respectable and decent, but not Earth-shattering. I'm guessing that bumping it up another 200MHz to get to the 3GHz mark would not be as beneficial as the initial 800MHz boost. Thoughts on this anyone?

 

[Epicenter]

So far you have proven nothing to me. It's true that cooling a CPU to oh, 100C below its normal operating temperature can increase stability with heavy overclocks. That's why LN2 works. But drops that are NOT as severe, won't yield such gains. So far nothing has been proven and that includes the image of the FX-55 at 3.46 GHz, with no mention of what was cooling it and a temperature monitor reporting 230C. What's that prove, that the video memory holds its contents when the CPU is on fire?

 

[GodsMadClown]

I'm guessing that bumping it up another 200MHz to get to the 3GHz mark would not be as beneficial as the initial 800MHz boost. Thoughts on this anyone?

If you're folding, then every bit counts.

 

[robberbaron]

So far you have proven nothing to me. It's true that cooling a CPU to oh, 100C below its normal operating temperature can increase stability with heavy overclocks. That's why LN2 works. But drops that are NOT as severe, won't yield such gains. So far nothing has been proven and that includes the image of the FX-55 at 3.46 GHz, with no mention of what was cooling it and a temperature monitor reporting 230C. What's that prove, that the video memory holds its contents when the CPU is on fire?

Well, when I had a Zalman CNPS7000A, 1.75v and 2.7ghz got me temps of 57C in Prime. It failed prime. When I moved to an XP90C, my load temps in prime in such settings dropped to 48C and prime was then stable. That's not too big of a drop, but I did manage to get the computer stable at a faster speed. The fact that lower temperatures yielded stabler higher speeds is enough evidence for me.

But I guess physics isnt your forte. It isn't mine either, but facts are facts.

 

[Epicenter]

Well, when I had a Zalman CNPS7000A, 1.75v and 2.7ghz got me temps of 57C in Prime. It failed prime. When I moved to an XP90C, my load temps in prime in such settings dropped to 48C and prime was then stable. That's not too big of a drop, but I did manage to get the computer stable at a faster speed. The fact that lower temperatures yielded stabler higher speeds is enough evidence for me.

But I guess physics isnt your forte. It isn't mine either, but facts are facts.

That isn't a big enough drop to impact performance. What I suspect is your CPU was really hotter than your sensor was telling you (in-socket perhaps?) and it was really running close to spec. Pulling it back into spec made it run stable in Prime95. 57C is simply not hot enough to cause heat related crashes and a ~9C temperature drop is not enough to allow a more stable overclock WITHIN operational temperature spec, simple as that.

While it's fun to say you have lower temps and that they are causing more stability, it simply isn't true. I admit I like to look at my CPU/GPU temp and see lower temps but it doesn't really do me any good. I have friends running CPUs within 10c of their max spec with the same stability I have FAR within it. So ... it's fun, but it's not factual.

 

[computerpro3]

So far you have proven nothing to me. It's true that cooling a CPU to oh, 100C below its normal operating temperature can increase stability with heavy overclocks. That's why LN2 works. But drops that are NOT as severe, won't yield such gains. So far nothing has been proven and that includes the image of the FX-55 at 3.46 GHz, with no mention of what was cooling it and a temperature monitor reporting 230C. What's that prove, that the video memory holds its contents when the CPU is on fire?

Okay, my a64 3500 on unmodded phase change was at -2C to +7C under load and did 3ghz as opposed to 2.6 on stock cooling.

What now?

 

[centvalny]

While it's fun to say you have lower temps and that they are causing more stability, it simply isn't true. I admit I like to look at my CPU/GPU temp and see lower temps but it doesn't really do me any good. I have friends running CPUs within 10c of their max spec with the same stability I have FAR within it. So ... it's fun, but it's not factual.

Just wait untill you get tired with your water cooling system....then you will see

 

[(cf)Eclipse]

Just wait untill you get tired with your water cooling system....then you will see

i agree. if what i have said before, and baron said isn't proof enough, before it took 1.57v to be stable at 2.85ghz. load temps was around 45c.
now i have water. load temp is 35c, and i'm priming at 2.86ghz with 1.55-1.54v. if you can find a different logical reason for the increase of max oc other than the lower temps, maybe you deserve some recognition.

until then, you've got 4 or 5 people against you and nobody with you

 

[DejaWiz]

This is an older article, but interesting nonetheless.

 

[Project_2501]

Man, this guy is stubborn. He doesn't realize that a better cooler can cool a hotter chip. The hotter chip is due to an increase in voltage, that was needed to OC.

here we go: my processor can't do 2.4 ghz 240x10 w/ a divider, at 1.5v, and the stock cooler isn't good enough to cool more than that. but it can do 230x10 at 1.5v.