Overclocking to 4ghz a cpu lifespan

camdenpaul

camdenpaul

camdenpaul

Associate
Joined
22 Dec 2010
Posts
9
Hey all, I wondering when you overclock a 3.06ghz i7 950 cpu to 4ghz how much of the lifespan is reduce on the cpu? also can someone here tell me if these bios settings are correct to get 4ghz that someone gave me, BCLK 200, Multi 20, dram freq 1600, pcie freq 101, load line calibration enabled or level 2, cpu vcore 1.3v or 1.325v, dram freq 1.65v and spread spectrums disabled, but i look in the bios and cant find spread spectrums anywhere, all i got is Isochronous support and virtualization technology is it one of them i gotta disable? my rig is 1000w psu, antec 902, constair a70 cooler, gigabyte ga-58-usb3 motherboard, memory Corsair XMS3 6GB (3x2GB) DDR3 PC3-12800C9 1600MHz Triple Channel, x 2 480 geforce cards
 
Sounds within the ballpark, with those volts in 'theory' degradation accleration will be minimal. Don't worry about the spread spectrum, never had that prevent stability. Whether the chip will be stable at those volts is totally up to luck but most are imo.
 
seems like you might just have a bit of a crappy chip if you need 1.3+ volts for 4GHz. i thought my chip was bad in that i needed that amount of voltage without LLC; with LLC i need 1.25 (obviously there are much better chips that can run 4GHz with under 1.2v).

as for life expectancy, you'll probably reduce its lifespan from 10+ years to 8 years. absolutely nothing to worry about
 
My personal 920 D0 needs 1.3v without LLC and 1.275v with LLC for 4Ghz, i find a lot of chips need close or more than 1.3v with LLC on for stability there. Have had the odd one that did 4Ghz on low volts, say 1.2v LLC, 1.25v ish without but those are rare imo. I personally build my systems with LLC off, have read some not so good things about it and vdroop is there for a reason. Time will tell i guess.
 
i was always under the impression that with older boards (s775 and older) that you were likely to suffer from transient spikes in voltage ~2v (but for a pico second), but with newer boards this isn't an issue due to the higher quality capacitors used.

i went through a time of not using LLC thinking it is probably safer, but i don't worry about it anymore, as ultimately by the time the chip is a dodo, i'd have upgraded.
 
Call me stubborn :D. With some stuff i'm like those ancient geezers who worship Intel, claim everything else is a clone and even back when AMD were whuppin em they never stopped praising the PIV, set in my ways :p
 
I need to Overclock that chip really as the standard stock settings when using sli with x 2 480 cards i see hardly any different in framerates in fact its the same as if i was using one card at stock settings, also i got told never to use easytune overclocking software, is it really that bad? but thanks for confirming them settings are correct with the bios as i was a little nervous about it cos i dont want the pc to blow up or something lol
 
VID range just means the range of VID's the chips will ship with. that doesn't mean the max safe voltage.

iirc, it's 1.55v absolute max before degradation (0.1v more than c2q max voltage).

edit: look at table 2-6 here
 
Last edited:
Processors will degrade at any voltage. Including stock settings. The doping ions move around as electrons crash into them until too many are in the wrong place and functionality suffers.

It'll degrade faster with higher voltages (not just Vcore), and at higher temperatures.

There is absolutely no way to answer your question in the sense you're hoping for. You don't know what its lifespan would have been at stock speeds, so it makes very little sense to ask how long it'll live under different conditions.

If you stay within the crowd with regards voltages, you'll probably be fine. If your particular chip is more fragile than the average, it wont be.
 
twist3d0n3 said:
VID range just means the range of VID's the chips will ship with. that doesn't mean the max safe voltage.

iirc, it's 1.55v absolute max before degradation (0.1v more than c2q max voltage).

edit: look at table 2-6 here
Click to expand...

Little misconception there. 1.55 is not absolute max working voltage, its the max it can cope with during a surge/spike etc..
 
i5 and i7, nothing about i3 ;)

that's what the datasheet says, so i'm inclinded to agree that that is the upper limit at what we should use. obviously, temps with voltages even above 1.35 cause massive amounts of heat - i doubt i could push mine to 4.4GHz without hitting 80*C even with a decent watercooling setup...

as for the surge/spike thing, i'll have to take your word for that then, as a lot of people on xs etc will use that as a general guide. i think i mentioned before regarding LLC, the ability to smooth out voltages with the capacitors in use on most modern motherboards is good enough to stop such transient spikes
 
In terms of lifespan 2 things will result in a large decrease in lifespan for the current material process for the i7 die >80C for prolonged periods and IIRC with the i7 upto 1.4v. Anything under 1.4v and <80C will not massively impact on life expectancy, as you go above 1.4v and 80C the impact on life expectancy ramps up massively. IIRC running a 45nm CPU at 1.4v @ 80C 24x7 should still last around 50,000 hours. Whereas say 1.45v @ 90C could bring it down to as little as say 5000 hours (of flat out running).

EDIT: Keeping temps down will help a bit but the voltage on its own has a massive impact and likewise keeping voltages low will help but high temps will kill it quite quickly too.
 
Last edited:
Nothing to do with the spikes occuring, just that 1.55v value is there as the upper limit. This does not mean that you are safe to run 1.55v constantly. As for 1.55v being absolute max before degradation, this is wrong as the cpu degrades no matter what the core voltage, electromigration is just sped up as voltages/heat goes up, at 1.55v this process will be tremendously higher.

It was at XS that i actually read about the upper limit myself, notice that the people at 1.55v usually only do it for benches/suicide runs and hardly 24/7? Also their overclocking is usually done with different methods of cooling where a chip might tolerate the voltages a little longer than on air, still dosen't mean it'll last long, look at the amount of dead cpus etc.. 1.55v is not meant to be a sustainable workable vcore.
 
Last edited:
CPUs last a very long time. Putting more Voltage through it does shorten the lifespan, although I'm not sure by how much. I think the general idea is that it's obsolete way before the damage becomes noticable.

With that said, I don't overclock at all - doesn't seem worth the risk.
 
Justintime said:
Nothing to do with the spikes occuring, just that 1.55v value is there as the upper limit. This does not mean that you are safe to run 1.55v constantly. As for 1.55v being absolute max before degradation, this is wrong as the cpu degrades no matter what the core voltage, electromigration is just sped up as voltages/heat goes up, at 1.55v this process will be tremendously higher.

It was at XS that i actually read about the upper limit myself, notice that the people at 1.55v usually only do it for benches/suicide runs and hardly 24/7? Also their overclocking is usually done with different methods of cooling where a chip might tolerate the voltages a little longer than on air, still dosen't mean it'll last long, look at the amount of dead cpus etc.. 1.55v is not meant to be a sustainable workable vcore.
Click to expand...

i didn't mean that's when it'll start to degrade, i phrased it poorly, i meant that it won't last long at all at that voltage, sorry for confusion :p

i completely agree with you, i'm not very comfortable about running above 1.4v tbh, i'm happy with 1.35v for 4.2GHz.

if i was going for a suicide run, i'd probably have a blast at that sort of voltage just to get a bench, but i'm not a bencher - i just appreciate the free performance ;)
 
billysielu said:
CPUs last a very long time. Putting more Voltage through it does shorten the lifespan, although I'm not sure by how much. I think the general idea is that it's obsolete way before the damage becomes noticable.

With that said, I don't overclock at all - doesn't seem worth the risk.
Click to expand...

it's definitely worth the risk of shortening its life by a few years, when you consider most of them last a helluva long time.

as you say, it'll more than likely become obsolete before it dies, so why not?

you're covered by a warranty if it does die in the first few years, and they have no way in hell of knowing if you've overclocked - as ultimately as jonj678 said, you have no way to tell how fragile the chip really is, and suffering the effects of electron migration can occur at any time in the chips life.
 
You must log in or register to reply here.
Back
Top Bottom