Opteron OverClocking voltage limits.

[themusgrat]

In your case, the 55mhz is not woth it. But I have talked to many people, and usually, a board's max volts will not kill a chip. Many times, after a period of high voltage, the chip will stop accepting it and only boot with low voltage, but most of the time, even 1.6 will not hurt it. And any chip will last at least 5 years, as a rule. For people who say otherwise, like 1 guy said 1-2 years, do you have first hand knowledge?No, they are guessing wrong.

 

[TrevorRC]

Originally posted by: sparkey
would you mind if i copied and pasted this quote to other forums that i belong to hav'nt yet but was tempted to last night
thanks tom


Originally posted by: TrevorRC
Know that you're running 80 Amps...
So; going from 1.4v -> 1.5v is equal to an extra 8 watts... going from 1.4 to 1.6 is an extra 16 watts. Puts the number into a bit better perspective than ".03 increase".

I personally try to NEVER run at 1.6 or higher, and I run water. The chips burn out much much faster if you put that much more voltage through them, and the warranty is gone when it dies--so if your extra 15% performance is worth the chip dying next year instead of 3 years from now, then by all means do it.

There is some rule of thumb that an extra 10Cs over the lifetime of a card/CPU cuts the life in half....maybe not quite that... but something along those lines.

Not at all; if someone else can confirm that what I said isn't absolute BS. (How embarassing would that be to have that quote everywhere and then have someone disprove it. Eh?)

I checked AMD's site to confirm that it was, in fact, 80 amps. But confirmation from someone with a multimeter would be nice

Also, if you can ammend it to say this:

Know that you're running 80 Amps...
So; going from 1.4v -> 1.5v is equal to an extra 8 watts... going from 1.4 to 1.6 is an extra 16 watts. Puts the number into a bit better perspective than ".03 increase".

I personally try to NEVER run at 1.6 or higher, and I run water. The chips burn out much much faster if you put that much more voltage through them, and the warranty is gone. So when it dies.. you're screwed. If your extra 15% performance is worth the chip dying next year instead of 3 years from now, then by all means go for it.

There is some rule of thumb that an extra 10Cs over the lifetime of a card/CPU cuts the life in half....maybe not quite that... but something along those lines.

 

[TrevorRC]

Originally posted by: themusgrat
In your case, the 55mhz is not woth it. But I have talked to many people, and usually, a board's max volts will not kill a chip. Many times, after a period of high voltage, the chip will stop accepting it and only boot with low voltage, but most of the time, even 1.6 will not hurt it. And any chip will last at least 5 years, as a rule. For people who say otherwise, like 1 guy said 1-2 years, do you have first hand knowledge?No, they are guessing wrong.

Guessing wrong?

Run 1.7 volts on any chip today, run it on air. You tell me how soon it dies.

Regardless of voltae, the silicon in the chip will degrade over time--with degredation comes a decrease in stability. The higher the voltage, the higher the heat, the faster the degredation. If you have higher voltage AND higher heat (higher voltage with no improved cooling) the lifetime degradation is squared.

 

[nealh]

There is some rule of thumb that an extra 10Cs over the lifetime of a card/CPU cuts the life in half....maybe not quite that... but something along those lines.

where did you obtain this data...so if you were to run your chip at 100% load 24/7(folding, seti etc) your temps are greater than 10C in general over idle...so will your chip life be shorter???

or are you talking 10C greater than max temps????? or are you saying an extra 10C over stock idle/load temps halves the life of your chip....

Hmm..10C with these chips is not alot...I would think there is more built in cushion ..different systems, components....

I personally try to NEVER run at 1.6 or higher, and I run water. The chips burn out much much faster if you put that much more voltage through them, and the warranty is gone. So when it dies.. you're screwed. If your extra 15% performance is worth the chip dying next year instead of 3 years from now, then by all means go for it.

TrevorRC..you make alot strong statements do you have facts to back them up..while I agree 1.6v is alot of vcore and should shorten cpu life nobody know for sure how much

There are several Intel engineers who have discussed this and consensus is it does and high vcore and higher temp combo is real bad vs high vcore alone or high temps alone

How do you know the chip will die in 1 yr vs 3yrs....

BTW you maybe correct but I would like to see some real data regarding this


I hope anyone who overclocks understands the risk of decreased life of the item

 

[TrevorRC]

Originally posted by: nealh

There is some rule of thumb that an extra 10Cs over the lifetime of a card/CPU cuts the life in half....maybe not quite that... but something along those lines.

where did you obtain this data...so if you were to run your chip at 100% load 24/7(folding, seti etc) your temps are greater than 10C in general over idle...so will your chip life be shorter???

or are you talking 10C greater than max temps????? or are you saying an extra 10C over stock idle/load temps halves the life of your chip....

Hmm..10C with these chips is not alot...I would think there is more built in cushion ..different systems, components....

I personally try to NEVER run at 1.6 or higher, and I run water. The chips burn out much much faster if you put that much more voltage through them, and the warranty is gone. So when it dies.. you're screwed. If your extra 15% performance is worth the chip dying next year instead of 3 years from now, then by all means go for it.

TrevorRC..you make alot strong statements do you have facts to back them up..while I agree 1.6v is alot of vcore and should shorten cpu life nobody know for sure how much

There are several Intel engineers who have discussed this and consensus is it does and high vcore and higher temp combo is real bad vs high vcore alone or high temps alone

How do you know the chip will die in 1 yr vs 3yrs....

BTW you maybe correct but I would like to see some real data regarding this


I hope anyone who overclocks understands the risk of decreased life of the item

If you want to give me ~$150 USD for a 144, I'll show you

Give me a bit to dredge some old stuff up.


Edit: Here we go:
http://www.overclockers.com/tips30/

CPU Life = Normal Life Hours / [((273 + New Temp) / (273 + Normal Temp)) ^ M]

So...
You do the math. :]

 

[nealh]

TrevorRC...

As I said you maybe 100% correct..I just want to see some info on this

I agree 100% 1.7v vcore should be a killer over time..1.6v...pretty high but with correct cooling maybe ok for awhile..how long who knows

 

[sparkey]

trev not a problem and very glad to amend it
this is for a guy that has the same board as i do he had his vc@1.475 with the over volt enabled (+.200) add it up and it = disaster over time
thanks again trev
tom

 

[TrevorRC]

Posted above.
Looking for CPU Lifespan hours. Assume ~10,000 Hrs/Yr.
I'll do the math in a bit, working on some other math right now (Calculus, believe it or not. Test tomorrow. Side note, if anyone on here is good with integration/surfaces of revolution/etc; please send me a PM/IM. Thanks.)
--Trevor

 

[fighterpilot]

So the bottom line is that even if you run at higher voltage as long as your temp is lower life of the chip isn't compromised as much?

I recently went to O/C my Opteron 175 and got it 2.7 GHz with 240FSB and 1.525vcore. I can get it it 2.6 with 1.475 vcore but then compromise memory speeds which dump down to 188Mhz since I have to change memory multiplier to 166MHz and HTT to 4x.

EDIT: Changed clock speeds from 3.7 and 3.6 to 2.7 and 2.6 respectively so as not to give other jealous overclockers a heart attack

 

[robertk2012]

Thats not the case. Even if the temps are kept low then a high enough voltage will fry it.

Originally posted by: fighterpilot
So the bottom line is that even if you run at higher voltage as long as your temp is lower life of the chip isn't compromised as much?

I recently went to O/C my Opteron 175 and got it 3.7 GHz with 240FSB and 1.525vcore. I can get it it 3.6 with 1.475 vcore but then compromise memory speeds which dump down to 188Mhz since I have to change memory multiplier to 166MHz and HTT to 4x

 

[fighterpilot]

Originally posted by: robertk2012
Thats not the case. Even if the temps are kept low then a high enough voltage will fry it.

So what's considered a "high enough voltage"???

 

[nealh]

Originally posted by: TrevorRC
Posted above.
Looking for CPU Lifespan hours. Assume ~10,000 Hrs/Yr.
I'll do the math in a bit, working on some other math right now (Calculus, believe it or not. Test tomorrow. Side note, if anyone on here is good with integration/surfaces of revolution/etc; please send me a PM/IM. Thanks.)
--Trevor

BTW 24hrs x 365= 8760 hrs

your link was excellent...but note since few of us run the an oced cpu at 100% all the time ..there is a good portion of time it is at idle..so the higher vcore has a lower effect on cpu life becasue heat is lower...

so higher vcore will shorten life but for most not to a point if they keep at a reasonable increase..10-15% IMHO

quote from the article
Estimating the Overclocking Impact on CPU Life

Now let's plug into this equation the impact on the C366's life between running the Running the CPU at spec and overclocking it. A heatsink of .35 c/w will result in a CPU temp of 32.6 C at spec for an expected life of 23,321 hours ("Average" curve). Overclocking the C 366 with the same heatsink will result in a CPU temp of 40 C resulting in an estimated life of 18,359 hours, a difference of 4,962 hours. Now what does this means? If you run this C 366 flat out for an average of four hours per day, running at spec will result in failure in 16 years. Running the same CPU overclocked will result in failure in 12.6 years. How long did you keep you last CPU?

of course this artice is 5+ yrs old so there maybe more of issue with higher vcore on new small die chips..who knows???

 

[robertk2012]

The only guarantee I can give you is that 110 volts from your wall socket would be enough. It all depends on your luck really. Each processor has its own tolerances so there is not a specific number. you might run a 165 at 1.8 volts for 10 years without a problem but someone else uses 1.5 and it does his processor in. I will tell you this that every increase in voltage increase your odds of failure exponetially....I read a post by some very bright minds one of which works at intel. Maybe you should do some searching on older posts on the same subject.

Originally posted by: fighterpilot

Originally posted by: robertk2012
Thats not the case. Even if the temps are kept low then a high enough voltage will fry it.

So what's considered a "high enough voltage"???

 

[fighterpilot]

Originally posted by: robertk2012
The only guarantee I can give you is that 110 volts from your wall socket would be enough. It all depends on your luck really. Each processor has its own tolerances so there is not a specific number. you might run a 165 at 1.8 volts for 10 years without a problem but someone else uses 1.5 and it does his processor in. I will tell you this that every increase in voltage increase your odds of failure exponetially....I read a post by some very bright minds one of which works at intel. Maybe you should do some searching on older posts on the same subject.

I notice in your sig that you have a 165 @ 2.889. What is your voltage, and are you using air cooling?

I'm thinking that I should have opted for the 165 and o/c'd that as most people tend to get to similar speeds as the 175. D'oh!

 

[robertk2012]

Its running at 1.4 volts air cooled. Its at idle right now at 31 degrees. At a full load the highest it has been at was 51 and thats before I had all the fans hooked up and moved some cables around to help airflow. through the case.

I could go higher but im limited now by the motherboard. If there was only a way to change the multiplier I cant complain though.

Originally posted by: fighterpilot

Originally posted by: robertk2012
The only guarantee I can give you is that 110 volts from your wall socket would be enough. It all depends on your luck really. Each processor has its own tolerances so there is not a specific number. you might run a 165 at 1.8 volts for 10 years without a problem but someone else uses 1.5 and it does his processor in. I will tell you this that every increase in voltage increase your odds of failure exponetially....I read a post by some very bright minds one of which works at intel. Maybe you should do some searching on older posts on the same subject.

I notice in your sig that you have a 165 @ 2.889. What is your voltage, and are you using air cooling?

I'm thinking that I should have opted for the 165 and o/c'd that as most people tend to get to similar speeds as the 175. D'oh!

 

[fighterpilot]

Originally posted by: robertk2012
Its running at 1.4 volts air cooled. Its at idle right now at 31 degrees. At a full load the highest it has been at was 51 and thats before I had all the fans hooked up and moved some cables around to help airflow. through the case.

I could go higher but im limited now by the motherboard. If there was only a way to change the multiplier I cant complain though.

Wow, how'd you manage 2.9GHz with 1.4V? That seems outrageous! The 175 is essentially the same chip, right? I can only get 2.6 with 1.4v, and 2.7 with .155v.

I will probably stick with the 2.6GHz only because of the voltage. But even I idle at about 30 and max is about 45 under full load. It must be something to do with my system.

 

[fighterpilot]

Originally posted by: nealh

Originally posted by: TrevorRC
Posted above.
Looking for CPU Lifespan hours. Assume ~10,000 Hrs/Yr.
I'll do the math in a bit, working on some other math right now (Calculus, believe it or not. Test tomorrow. Side note, if anyone on here is good with integration/surfaces of revolution/etc; please send me a PM/IM. Thanks.)
--Trevor

BTW 24hrs x 365= 8760 hrs

your link was excellent...but note since few of us run the an oced cpu at 100% all the time ..there is a good portion of time it is at idle..so the higher vcore has a lower effect on cpu life becasue heat is lower...

so higher vcore will shorten life but for most not to a point if they keep at a reasonable increase..10-15% IMHO

quote from the article
Estimating the Overclocking Impact on CPU Life

Now let's plug into this equation the impact on the C366's life between running the Running the CPU at spec and overclocking it. A heatsink of .35 c/w will result in a CPU temp of 32.6 C at spec for an expected life of 23,321 hours ("Average" curve). Overclocking the C 366 with the same heatsink will result in a CPU temp of 40 C resulting in an estimated life of 18,359 hours, a difference of 4,962 hours. Now what does this means? If you run this C 366 flat out for an average of four hours per day, running at spec will result in failure in 16 years. Running the same CPU overclocked will result in failure in 12.6 years. How long did you keep you last CPU?

of course this artice is 5+ yrs old so there maybe more of issue with higher vcore on new small die chips..who knows???

So, I'm getting confused here. I am hearing that high voltage is bad because of the heat, so make sure you have proper cooling. Well, my CPU even at 1.55vCore doesn't exceed 45C, which is great IMHO, and only a few degC higher than at stock speeds. If it exceeded 55C then I'd start to be concerned. So should I not be concerned because it only bumps the temps by a few degrees?

And also, what is stock voltage for the Opteron 165/175? Is it 1.3 or 1.35? If it's 1.35 then 15% increase in voltage is about 1.55 volts, so it would be on the edge of the "acceptable" limits, but 10% says not to exceed 1.5v.

Either way, if temp decreases the life from 16 to 12 years, and added 15% volteage halves it, then I'm not concerned, that puts it at 6 years. Hell, even at 3 years, I wouldn't worry. It's if it blows out at less than a year that I am really concerned.

Guess I'll just run at a happy medium of 1.50 volts and 2.65GHz. An extra 50 MHz probably means 1 FPS in a game. I'm not going to conern myself with that if it means risking a premature failure.

 

[Syndicate]

I ran a AXP 1700+ DLT3C (stock voltage 1.45) at 2.3ghz and 1.89v for about 4 years. Air cooling that kept below 50c at prime 95. It never died and is still going strong in another box. I'm currently running a Opty 144 @ 2.8ghz with 1.65v air cooling. Idle temp is 30c, load temp is 46c.

 

[fighterpilot]

Originally posted by: Syndicate
I ran a AXP 1700+ DLT3C (stock voltage 1.45) at 2.3ghz and 1.89v for about 4 years. Air cooling that kept below 50c at prime 95. It never died and is still going strong in another box. I'm currently running a Opty 144 @ 2.8ghz with 1.65v air cooling. Idle temp is 30c, load temp is 46c.

Sounds promising...

 

[robertk2012]

set at 1.4 in the bios. DFIs have a tendency to overvolt a bit so its probably a bit higher. either way. I can hit around 2.6 on stock volts

Originally posted by: fighterpilot

Originally posted by: robertk2012
Its running at 1.4 volts air cooled. Its at idle right now at 31 degrees. At a full load the highest it has been at was 51 and thats before I had all the fans hooked up and moved some cables around to help airflow. through the case.

I could go higher but im limited now by the motherboard. If there was only a way to change the multiplier I cant complain though.

Wow, how'd you manage 2.9GHz with 1.4V? That seems outrageous! The 175 is essentially the same chip, right? I can only get 2.6 with 1.4v, and 2.7 with .155v.

I will probably stick with the 2.6GHz only because of the voltage. But even I idle at about 30 and max is about 45 under full load. It must be something to do with my system.

 

[kravmaga]

I have my opty 2750mhz at 1.52. I took it up to 1.55 and tried for 2.8 but it just wasn't stable. My temps however are extremely good considering and in contrast to my single core A64 CPU at the same voltage. I'm at 32 idle and I doubt I even break 45 with full load. If I could go a bit higher I think I could get 2.8+ stable but I don't know how safe that is. I'm using an XP-90c with medium speed panaflo fan..

 

[MTDEW]

Originally posted by: fighterpilot
So the bottom line is that even if you run at higher voltage as long as your temp is lower life of the chip isn't compromised as much?

I recently went to O/C my Opteron 175 and got it 3.7 GHz with 240FSB and 1.525vcore. I can get it it 3.6 with 1.475 vcore but then compromise memory speeds which dump down to 188Mhz since I have to change memory multiplier to 166MHz and HTT to 4x

3.6 - 3.7ghz!!
Typo?

If not, PM ME , i have a wife and several extra limbs i'd be willing to part with for that thing.

 

[fighterpilot]

Originally posted by: MTDEW
[

3.6 - 3.7ghz!!
Typo?

If not, PM ME , i have a wife and several extra limbs i'd be willing to part with for that thing.

Edited and fixed. Sorry about that. Don't need another wife either, thanks! Current one is more that I can handle. You can have mine for free!