BF2S Forums A 56€ CPU and me

Lucien wrote:

GC_PaNzerFIN wrote:

1.5675v

eh?

Lucien wrote:

GC_PaNzerFIN wrote:

Lucien wrote:

eh?

why not?

From what I've read on forums that's a dangerously high voltage, no?

GC_PaNzerFIN wrote:

_NL_Lt.EngineerFox wrote:

70c , thats fucking shortens the life right out of it, I keep my 45nm CPU at 55c Load on 3.6Ghz with Boxed Cooler.

nah. 70c won't shorten its life. as long as it stays under 75c it is good. the 45nm CPUs need to be under 70c tho.

_NL_Lt.EngineerFox wrote:

gtfo with your prehistoric CPU

Aries_37 wrote:

Nice, for 50 bucks that's a lot of performance. But you are pushing it a bit hard though
70C and 1.56V+ when with speedstep it runs under 50C and ~1.28V stock is just a bit too close to tjunction..

DeathUnlimited wrote:

Only bad thing about it is the low cache. I love it too, at 3.0GHz (1.4 vcore) with E6300 boxed cooler. Tops at 69C during 3dmark06 CPU test

I think you might have got got a bit unlucky with your chip or you're unnecessarily overvolting. Did you jump straight to 1.4V? I OCed my parents' e2200 for fun; dropped it's multiplier to 10 (so its basically an 2180) and got it to 3ghz at 1.34V and 55C under load with a cheap freezer7pro. Is it really not stable at any lower vcore?

Last edited by DeathUnlimited (2008-05-02 11:51:43)

Bertster7 wrote:

GC_PaNzerFIN wrote:

_NL_Lt.EngineerFox wrote:

70c , thats fucking shortens the life right out of it, I keep my 45nm CPU at 55c Load on 3.6Ghz with Boxed Cooler.

nah. 70c won't shorten its life. as long as it stays under 75c it is good. the 45nm CPUs need to be under 70c tho.

70C WILL shorten its life substantially. It's not hot enough for it to imminently break, but it's too hot. Anything over 60C at any point is damaging for the CPU (based on 65nm Conroe architecture at least).

70C is pushing the boundaries of acceptable temps. It is too hot, but only just. It's fairly unlikely that the CPU won't last 3 or 4 years running like that. The problem is that if your ambient temps increase by 5-10C which in summer they often do, then your CPU becomes quite dangerously hot.

Bertster7 wrote:

Freezer7Pro wrote:

Bertster7 wrote:

70C WILL shorten its life substantially. It's not hot enough for it to imminently break, but it's too hot. Anything over 60C at any point is damaging for the CPU (based on 65nm Conroe architecture at least).

70C is pushing the boundaries of acceptable temps. It is too hot, but only just. It's fairly unlikely that the CPU won't last 3 or 4 years running like that. The problem is that if your ambient temps increase by 5-10C which in summer they often do, then your CPU becomes quite dangerously hot.

Erm, no.

Yes....

60C* is the point at which the rate the silicon becomes degraded increases rapidly (in fact I think they currently use strained silicon, which came in with the San Diego core AMD CPUs, which works off an SiGe substrate layer to increase carrier mobility by making each carrier further apart). Prolonged usage at 80C* will often kill a CPU. When OCing 60C load temps is where you should stop if you want to be confident that your CPU will last a good few years.
Neither of these temps are going to cause hot electron degradation of the FET gates straight away, but they are temps around the point where this becomes possible and prolonged usage at these temps may cause the CPU to fail, particularly when the silicon is in a weakened state from running hot for a long time. You've also got to consider the increased resistance hotter CPUs provide and the negative effects this can have on them - though these temps are not really high enough for that to be a major factor.

Of course there are loads of cases where this doesn't happen because each CPU is different. But as an average, it is very true. Of course if you're spending very little on a CPU and OCing it to achieve massively improved performance, then a probable, significant decrease in your CPU life is not a big factor when it comes to getting bang for your buck.

It's worth noting you should be much more careful when OCing the new 45nm CPUs. Since they break a lot easier due to the manufacturing process they use....

*It's also worth pointing out that it isn't just the heat that cause the degradation. It is a combination of the heat, voltage and frequency (people always seem to assume frequency has no impact on degrading the CPU, they're wrong). Those temps are fairly accurate for running at 1.5V/3GHz on a Conroe.

Bertster7 wrote:

Freezer7Pro wrote:

Bertster7 wrote:

Yes....

60C* is the point at which the rate the silicon becomes degraded increases rapidly (in fact I think they currently use strained silicon, which came in with the San Diego core AMD CPUs, which works off an SiGe substrate layer to increase carrier mobility by making each carrier further apart). Prolonged usage at 80C* will often kill a CPU. When OCing 60C load temps is where you should stop if you want to be confident that your CPU will last a good few years.
Neither of these temps are going to cause hot electron degradation of the FET gates straight away, but they are temps around the point where this becomes possible and prolonged usage at these temps may cause the CPU to fail, particularly when the silicon is in a weakened state from running hot for a long time. You've also got to consider the increased resistance hotter CPUs provide and the negative effects this can have on them - though these temps are not really high enough for that to be a major factor.

Of course there are loads of cases where this doesn't happen because each CPU is different. But as an average, it is very true. Of course if you're spending very little on a CPU and OCing it to achieve massively improved performance, then a probable, significant decrease in your CPU life is not a big factor when it comes to getting bang for your buck.

It's worth noting you should be much more careful when OCing the new 45nm CPUs. Since they break a lot easier due to the manufacturing process they use....

*It's also worth pointing out that it isn't just the heat that cause the degradation. It is a combination of the heat, voltage and frequency (people always seem to assume frequency has no impact on degrading the CPU, they're wrong). Those temps are fairly accurate for running at 1.5V/3GHz on a Conroe.

Well, that's the theory, at least. If it's true remains to see. I've seen numbers lie before, and I won't be surprised if they do again. We'll see in a couple of years.

I see it in practice every day.

Doing repairs for Apple gives me plenty of opportunity to deal with Conroe based systems with totally inadequate cooling. iMac CPUs die every 5 minutes. Then they need new logic boards (since the CPU is integrated).

Last edited by GC_PaNzerFIN (2008-05-03 07:42:07)