Death by OC, share your experiences

[blackened23]

But I had my 17 875K on the predecessor of adaptive Voltage -- the Offset Voltage. It still lost its OC.

As I said nothing is guaranteed and nothing is fail safe. If you over-volt be prepared to lose the chip. It'll take a while but it can happen no matter what. I never said otherwise. Usually takes 1.5+ years, my lynnfields went after around 2 years of OV'ing.

But the chances of losing it are lessened by a lot when using variable voltage.

 

[AznAnarchy99]

Had 2 i7-870s run for about 1.5-2 years in separate systems give out the same way. They were OC'ed with manual voltage 24/7 and eventually both systems started acting up at OC'ed settings. Then was normal at stock. Then stock started acting up. Then every clockspeed started acting up.

Death by over-voltage electromigration. Although in hindsight I should have used variable voltage, but that's no guarantee of a shorter lifespan.

Oh and, I know the CPUs gave up from over-volting too long because I replaced both with a drop in replacement and both systems worked fine afterwards with a new CPU. Nowadays I *never* use manual voltage for over-volt OC'ing for this very reason, it takes a few years but will increase the rate of electromigration if you're over-volting. Lesson learned. I use manual voltage to dial it in, but after that I always use adaptive or offset voltage when over-volting a CPU. Not a guarantee but definitely helps increase the lifespan. Now i'm aware that some have over-volted for 5-10 years with NP, so have I with older CPUs, but I wasn't so lucky in these two situations.

D:

I've been running my i7-860 like this since 2009. Well I guess if it craps it's due for a replacement anyways.

 

[YBS1]

Not so far, I'll let you know after a little more time with my X5670. I guess where it's not my main pc and it was just a cheap toy to play with I've been more willing to try stupid stuff with it. In fact a couple of days ago I had an oopsie with it. While still trying to get it benchmark stable at 5025MHz I had tried a few new things with the settings. I run with LLC enabled for my 24/7 settings but when I'm really pushing the voltage/OC I disable it so it won't climb above the voltage I set. In my haste I had left LLC enabled....I fired up Cinebench R15 and as it started running I noticed the voltage a bit above what I would have liked and thought to myself, "That's a bit odd". A micro-moment later the voltage jumped up to 1.73....My thought changed from "a bit odd" to more of a Ralphie spills the lug nuts moment. Shut that down quick.

On the flip side of the argument, I swear my W3520 actually got stronger the longer it was overclocked. The longer I had it the faster it went. lol

 

[WhoBeDaPlaya]

D:
I've been running my i7-860 like this since 2009. Well I guess if it craps it's due for a replacement anyways.

I think you should be fine. My two i7 920 D0s @ 4.1GHz have been running DC pretty much 24/7 since they were built in 2009.
Recently revalidated the OCs (IBT, Prime95, x264 stability v2) - passed with flying colors.

 

[Bubbleawsome]

On the flip side of the argument, I swear my W3520 actually got stronger the longer it was overclocked. The longer I had it the faster it went. lol

Hm, that's odd. Should've kept it, might have hit 10ghz by 2020.

D:

I've been running my i7-860 like this since 2009. Well I guess if it craps it's due for a replacement anyways.

It'll probably be ok. My 870 ran at 1.3v stock and the 45nm process could take a beating in the later revisions. I've seen some up at 1.5v still going.

 

[tolis626]

Well, to be honest, I take the fact that we have next to no data regarding long-term damage of OC/OV Haswell CPUs (or Ivy Bridge for that matter) as a good thing, which means there aren't many chips, if any, that have started to fail prematurely when overclocked to sane speeds. Now if someone overclocks their chip using manual voltage over 1.4V 24/7... That's another story.

 

[Ho72]

Well, to be honest, I take the fact that we have next to no data regarding long-term damage of OC/OV Haswell CPUs (or Ivy Bridge for that matter) as a good thing, ...

I think it's a little premature to call the data we have at this point long term.

As I said, I tend to run my rigs 5 years at a minimum, so the data points I'm interested in with the recent generation of processors have yet to be plotted.

 

[tolis626]

I think it's a little premature to call the data we have at this point long term.

As I said, I tend to run my rigs 5 years at a minimum, so the data points I'm interested in with the recent generation of processors have yet to be plotted.

I said long-term damage, not data.

Although now that I think of it, "long-term" might not have been the best wording I could choose. :hmm:

What I meant was that it's been a little over a year after the first Haswells hit, and we have yet to see data regarding damaged chips with any statistical meaning. Sure, you'll get a fried chip here and there, but defective ones and user error should more than cover these two cases. I haven't even heard of chips degrading, and degradation can happen rather quickly in some cases.

To be clear, I'm not at all saying this is somehow conclusive. It's just that the abscence of such data makes me hope that Haswell and Ivy Bridge chips will hold out well enough.

 

[Ho72]

To be clear, I'm not at all saying this is somehow conclusive. It's just that the abscence of such data makes me hope that Haswell and Ivy Bridge chips will hold out well enough.

Me too.

 

[GLeeM]

I'm more interested in how common it is for a stable, 24/7 overclock to have issues down the road, say past the 3 year mark.

My two i7 920 D0s @ 4.1GHz have been running DC pretty much 24/7 since they were built in 2009.

No issues yet!
i7 920 D0 @ 3.95GHz @ 1.264v on which I too run DC, so far ~45,000 hours 100% load with temps 75-85C (TRUE120).

It might be that I don't have any problem because it is always at 100% constant load and not up and down?

My current Real Temp reading is:
281:22:50 hours:minutes:seconds
Current temps 76 75 75 74
Min temps 39 40 39 37
Max temps 85 85 84 82

 

[tolis626]

No issues yet!
i7 920 D0 @ 3.95GHz @ 1.264v on which I too run DC, so far ~45,000 hours 100% load with temps 75-85C (TRUE120).

It might be that I don't have any problem because it is always at 100% constant load and not up and down?

My current Real Temp reading is:
281:22:50 hours:minutes:seconds
Current temps 76 75 75 74
Min temps 39 40 39 37
Max temps 85 85 84 82

Holy crap, that's a lot of load. Why do you torture the poor thing?

Unfortunately, we can't use this as an example of how a Haswell or Ivy CPU will age while overclocked. The 920 was what? 45nm? For all we know, 22nm FinFETs may be completely different. Current CPUs may just start dying suddenly within the first two years, or they may live on after they have outlived their usefullness and then some. That and, well, I guess even for a 920 1.264V isn't all that much voltage. 45000 hours at 100% load is out of this world anyway, so I guess that even 1.264V could create problems.

 

[tolis626]

Sorry, double post

 

[TeknoBug]

Back in the Athlon Slot A days I killed a motherboard but can't remember how, also had a motherboard fail with a Thunderbird 1.2GHzz. Never killed any of my Intel systems, including the dual Pentium Pro server with 6x SCSI drives in it generating enough heat to be a central heating unit for an apartment (and it ran for over 12 years).

 

[GLeeM]

Why do you torture the poor thing?

Running distributed computing programs is not torture, it is what the CPU is designed to do. Granted not the overclock part
I am more worried about the mobo giving out.

But yeah I would like to know if I could run my 4790K at 4.7 for 5 years without degrading. It would mean a lot more points/work done over being safe at 4.5GHz.

 

[Dravic]

Only failure to OC i can think of:

KIII 350 @ 500 - eventually wouldn't even run at 350 anymore..


my two biggest O'C'.s % wise just became obsolete tech...

Cel 333 @ 500 - in a box somewhere
Duron 650 @ 1000mhz - in a box somewhere

Lost a few cheap PSU's along the way, but no boards, and just the 1 CPU

 

[hoorah]

I had a Geforce2 MX200 with no heatsink on it (cards actually came that way that long ago). I used thermal adhesive to stick a heatsink on it and OCed the card a bunch, can't remember how fast. Got decent performance out of it.

At some point down the road the paste came unstuck and the heatsink fell off into the board. The card very quickly would lock up during gaming in a dazzling display of corrupted purple.

Found the issue, stuck the heatsink back on and it seemed to be okay at stock but couldn't take the OC anymore.

 

[BonzaiDuck]

Well, the question is sufficiently specific:

"have you ever significantly degraded or killed a CPU by overclocking . . "

And the answer for me is "no." Motherboards? At least one that I know of personally.

I won't say that my practice is the reason I've avoided degradation. I'm just a bit cautious: I stop over-clocking when the core temperatures exceed the processor TCASE spec by 10C, and I stop overclocking when the highest measured voltage exceeds what we "either know or deduce" to be the upper "safe-range" limit.

I don't even know if Intel posts a "TCASE" spec anymore, and in the last couple generations of processors, the consensus may have only deduced a "safe-voltage" limit from a chip's lithography.

But I haven't built a system for personal ("favored") usage since 2004 that I didn't overclock.

Just an afterthought, though. I HAVE killed a memory-module or two from that practice, maybe as recently as 2008. But if I'd been born then, I'd be playing little-league now . . . .

 

[sxr7171]

I feel like after 2 years I had to increase my voltage on my 3570k for a stable 4.8GHz. Just a minor increase. Possibly Win 8.1 finickiness is the cause.

 

[Puppies04]

My first 2500K degraded after about a year running at 4.6, it started throwing up blue screens. I bumped the CPU voltage a couple of notches and it ran fine, still does 2 years down the line, I sold it to a friend and it hasn't missed a beat. I guess I was just above the stable threshold on the first overclock and it just took a tiny amount of degradation to make it unstable.

 

[TeknoBug]

I forgot to mention that I once had a dual Celeron 533MHz @ 1.1GHz on an Abit board, never once hiccuped... it was an unbelievable mahine.

 

[BonzaiDuck]

My first 2500K degraded after about a year running at 4.6, it started throwing up blue screens. I bumped the CPU voltage a couple of notches and it ran fine, still does 2 years down the line, I sold it to a friend and it hasn't missed a beat. I guess I was just above the stable threshold on the first overclock and it just took a tiny amount of degradation to make it unstable.

I offer nothing solid here, but to say it was once the case that mobo components would degrade. Newer gen boards use solid-state components -- mosfets, capacitors, etc. It's still probably a good idea to give the system a "shakedown" every six months with IBT or LinX -- enough to assure that you can get through 10 or 20 iterations.

Also, YMMV -- depending on how you use the system. I've been running three-plus years now at similar speeds (see sig). My blue-screens were due to a combination of things -- but not the RAM, mobo or CPU. I had a UPS system (Cyber-Power) that wasn't compatible with my Seasonic Active-PFC PSU, and I think there was a conflict over USB3 and SATA resources. But the OC settings had nothing to do with it -- proven by running at stock settings.

What were your load voltage and "turbo-unloaded" voltage? That is, the voltage under maximum droop, and the higher (and highest) voltage shown at termination of a PRIM95, IBT or LinX test?

 

[Puppies04]

What were your load voltage and "turbo-unloaded" voltage? That is, the voltage under maximum droop, and the higher (and highest) voltage shown at termination of a PRIM95, IBT or LinX test?

Too long ago to remember tbh. I am going round to upgrade his brothers GPU soon though so I might take a look.

 

[Deders]

I feel like after 2 years I had to increase my voltage on my 3570k for a stable 4.8GHz. Just a minor increase. Possibly Win 8.1 finickiness is the cause.

My first 2500K degraded after about a year running at 4.6, it started throwing up blue screens. I bumped the CPU voltage a couple of notches and it ran fine, still does 2 years down the line, I sold it to a friend and it hasn't missed a beat. I guess I was just above the stable threshold on the first overclock and it just took a tiny amount of degradation to make it unstable.

Same with my i5, after a while I noticed one core wasn't as stable under prime64 as it had been previously. I had to raise the VTT a notch or 2 but it's been fine ever since (touch wood) This may have been after I RMA'd my ram because it didn't work at full speed.

Been going strong for 5 years now, hope I don't jinx it :\

 

[YBS1]

I answered no earlier, but it just occurred to me that I quite possibly did a long time ago though I can't say for certain if it was the CPU or motherboard. It was a 1333 Thunderbird chip, pretty decent overclock IIRC. Just went to power it on one day and got nothing but spinning fans. I didn't investigate it too hard as I used it as an excuse to upgrade to an Athlon XP.

 

[njdevilsfan87]

I remember pushing a golden Q6600 to 3.8ghz on some lower end Gigabyte board. About 30 seconds into the stress test there was a very foul smell that came out of the case. I shut everything down immediately... and everything still worked fine without any visible damage. That was probably the closest I've come to frying anything. I still own the Q6600 (over 7 years old now) and it can still put 3.6ghz overclocks on a P35-DQ6 board.