OC will shorten the hardware life span?

[wilson0416]

Hi, I'm newbie in OC. My inquiry as stated in the title above.

 

[Idontcare]

Answer: Yes.

 

[Smoblikat]

Not necessarily, extra voltage will. But an undervolt combined with an overclock might break even/extend the lifespan.

 

[LOL_Wut_Axel]

Only by a small amount if you're overclocking with no overvolting. If you're overvolting lifespan will decrease exponentially, though a moderate overvolt of say 1.25-1.3V on Sandy Bridge or Ivy Bridge should mean it'll have no problems running for at least five years.

 

[Janus67]

Technically yes, but under most circumstances and hardware life cycles you will have likely replaced the hardware by the time the life span had decreased enough to be noted.

 

[wilson0416]

Thank you so much for answering my inquiry.

 

[Rvenger]

It's really hard to determine an actual CPU lifespan. For all you know it could have lasted the same amount of time at stock vs overclocked. I still know people with overclocked computers(Athlon XP ) that are going on 10 years old and still kicking overclocked like day 1. It's a toss up really... I guess its all part of the silicon lottery.

 

[HURRIC4NE]

what if you're cooling the hell out of the cpu? does that help apart from keeping the cpu cooler?

 

[Soccerman06]

what if you're cooling the hell out of the cpu? does that help apart from keeping the cpu cooler?

No, voltage is what kills a cpu, only preventing the cpu from hitting its max temp will prevent damage.

 

[Idontcare]

No, voltage is what kills a cpu, only preventing the cpu from hitting its max temp will prevent damage.

Thermally activated degradation mechanisms in IC's follow the Arrhenius equation.

In practice this means the lifetime of an IC roughly doubles for every 10°C decrease in operating temperature if all else is held constant (voltage, clockspeed, workload, humidity, etc).

Degradation mechanisms for an IC are dependent on voltage (electric field gradient), current (clockspeed, workload), and temperature (activation barrier).

 

[wilson0416]

Is there any method to maintain the hardware life span? I'm using q9550 & wish to overclock to 3.2Ghz. I'm totally new to OC.

 

[Idontcare]

Is there any method to maintain the hardware life span? I'm using q9550 & wish to overclock to 3.2Ghz. I'm totally new to OC.

You can buy a better heatsink cooler for your OC'ed CPU, lower its temperatures and improve the life span.

Lowering the peak temperatures also has the secondary effect of lowering the voltage required to maintain stability.

So - lower the temperatures and you improve lifespan. Take advantage of the lower temperatures to further lower your voltage and you improve lifespan. And both improve the power consumption (lowers it) which then costs you less money over the years.

Here is an example of the real-world results with an Ivy Bridge CPU, the same idea applies to your Q9550 as well:

^ lowering the operating temperatures by 20°C enabled me to further lower the operating voltage by nearly 0.04V, which in turn enabled a further 3°C lowering in operating temperatures. All of which improved the life span for the hardware while also lowering its power-consumption.

 

[LOL_Wut_Axel]

Is there any method to maintain the hardware life span? I'm using q9550 & wish to overclock to 3.2Ghz. I'm totally new to OC.

Use the least amount of voltage possible to get the system to be completely stable. I recommend you download IntelBurnTest, run it at the High preset use that for stability testing. 3.2GHz is incredibly easy to achieve with a Q9550, though. I'm 99% sure you'll be able to reach that with stock voltage. Just raise the FSB on your BIOS to 380MHz, leave it at stock voltage, and run IBT.

Of course, it should go without saying that if you're gonna overclock you should use a different heatsink than the stock one. Buy something like a Cooler Master Hyper 212+ or a Corsair A50.

 

[2is]

Not necessarily, extra voltage will. But an undervolt combined with an overclock might break even/extend the lifespan.

It's also a great way to introduce random instability.

My OC mentality is to find the highest stable OC possible with the lowest voltage that still keep me well within thermal limits. Then I drop the multi by 1 to further ensure stability.

 

[blastingcap]

Hi, I'm newbie in OC. My inquiry as stated in the title above.

Here, read this:

http://www.anandtech.com/show/2468/6

It's about CPUs, but applies to other silicon circuits as well.

 

[VirtualLarry]

It's also a great way to introduce random instability.

My OC mentality is to find the highest stable OC possible with the lowest voltage that still keep me well within thermal limits. Then I drop the multi by 1 to further ensure stability.

I wouldn't drop the multi on a Gigabyte 775 board. I tried that with my E2140s, and while they were stable at 8x400, they were NOT stable at 6x or 7x multi. Wierd but true. So now, I only run at max multi on Gigabyte mobos.

 

[pauldun170]

Is there any method to maintain the hardware life span? I'm using q9550 & wish to overclock to 3.2Ghz. I'm totally new to OC.

I have a Q9550 in a ep45-ud3p running at 3.4ghz.
To overclock, all I did was raise the front side bus. I also undervolted it to 1.17.

Its cooled with a cm 212+ bought on sale for someteen dollars and pocket change.

 

[ShintaiDK]

Never OC if you cant afford to lose your HW. Plain simple.

 

[Idontcare]

Never OC if you cant afford to lose your HW. Plain simple.

True that. And I'd extend this sage advice to include "Never OC if you can't afford to lose your data (files)."

Silent data corruption really sucks.

Undetected- also known as "silent corruption". These problems have been attributed to errors during the write process to disk. These are the most dangerous errors as there is no indication that the data is incorrect.

 

[Ferzerp]

True that. And I'd extend this sage advice to include "Never OC if you can't afford to lose your data (files)."

Silent data corruption really sucks.

I actually managed to scramble quite a bit of data last week when OC'ing my video card...... I have an SRT set up, so I have a lot more points of failure in my storage system. The system wouldn't boot, and when I finally repaired that, half my applications didn't work anymore, etc.

I always run a non-oc'ed WHS server (now with WD RED drives), so I always have a roll-back. All I did that day was a bare-metal restore on the previous day's image. (It takes a few hours, but that's ok)

 

[Idontcare]

Yep, thankfully you know what you doing and made plans in advance to mitigate the risk of data corruption. Not too many folks know to even think about it, let alone know what they can do to minimize the risk of it.

 

[Ferzerp]

It's worth noting, that if random, undetected corruption existed, I'd likely not have a good backup because it would lilkely age out before I detected it.

I don't store important data on the desktop though. Typically, anything important is accessed with the laptop, and stored on the server. The desktop is a gaming/tinkering machine primarily.

 

[shortylickens]

I cant remember the last time I kept a CPU until it burned out.

I think my 386 still works.

 

[KingFatty]

Are CPU's even provided with a lifetime estimate, such as a mean-time-before-failure (MTBF) that you'd see on a hard drive?

I mean, if the manufacturer won't even feel the need to let you know about a lifetime expectation, doesn't that say something about the nature of CPUs and that you just shouldn't even begin to worry about their lifetimes?

 

[BonzaiDuck]

I've had machines crap out in less than six years, never over-clocked, because of motherboard or PSU failure.

I've had systems overclocked from day 1 running 24/7 longer than five years. But always, from day 1, I'm trying to keep within a pre-chosen voltage limit based on all the information I can find on the processor.

I've seen people on various forums go into panic because they were too enthusiastic and impatient with their OC projects -- with failed processors, processors that were "misbehaving," and other issues. Always -- the underlying scenario involved pushing the VCORE voltage way beyond practical recommendations, failure to pay attention to thermal specs and limits, etc.

In my own little world, I may see myself as a "Walter White" [AMC's "Breaking Bad"] of OC'ing. Ten or fifteen years ago, I would tell you that OC'ing was irresponsible, shouldn't be done on machines with any serious usage or data [like "better living through chemistry" excluded becoming a "cook."] Now, I over-clock everything except my WHS server. There isn't any law against it, but I willfully break my warranty before I even install the processor.

To deal with data-loss, I back up regularly three different ways -- all part of a cycle. I clone my Oc'd system's hard-drive; I back up or save directly important files to the server; and then I back up the server.

Truth be told, I've never lost data due to corruption; I've only lost "misplaced or forgotten data."