Share your degradation reports on the 2500K/2600K

[Blackbird_CaD_]

This thread baffles me.
Before i read this thread i thought that under 1.4 is safe for 24/7.
But in this thread 1 guy killed his CPU after 2 months @ 1.40, 1 guy got degradation after 3 months @ 1.36, and 1 guy got degradation in less than 1 month @ 1.38????
I wish i had never seen this thread, i don't know if i dare to overclock at all anymore.

[PiLsY]

3 dead cpus out of the countless sandybridge owners who have read this thread? And all 3 at "safe" voltages that the vast majority of users on here run at if not higher for the last year or so?

Vcore was not the root cause for these cpus degrading.

Other factors must have been involved to cause these 3 cases of cpu degradation. Also note there are only 3 reports of cpu degradation on the forum out of many hundreds if not thousands of sandybridge overclockers viewing or posting here. There are too many people running 1.5v or more still having no issues at all for vcore to be an issue. Keep your temps under control and you're good.

[QuadDamage]

I'm still at 5 Gee's 1.5V load, 1.515V idle. Been running for months, or a year now and that's with a really underwhelming Gigabyte mobo. I also need 1.825V PLL anyhting less will cause idle BSODs. According to my MM it actually spikes up to 1.85V as soon as i fire up a game or whatever. I'm pumping around 1.125V idle/1.135V load (1.10V in BIOS) thru the IMC as well. All of my recent chips ie 940/980X ES, 990X, 920 D0's they all run at atleast 1.5V for a long time and they're still alive and just as good as they were the day I got them.

Those dead cpu's are prolly the victims of bad motherboards and/or PSU's. Like flaky rails, bad caps, bad LLC implementation etc.

[tcclaviger]

Not a 2600, but similar

CPU: 3820
Vcore(bios): Normal using +0.54v offset, LLC Medium
Vcore(DMM): 1.528v loaded
VTT: 1.038v
vDIMM: 1.7v
PLL: 1.703v
Mobo: Rampage IV Gene
Degredation: After the 1st day I had to bump voltage up .02 higher than the very first OCs, other than that none in a month so far.

Comments:
Chip runs this 24/7 crunching Rosetta@Home which loads it to about 70c on the hot cores and 66c on the cooler ones. The only less than 100% load it ever sees is when I am using it, then it's typically between 50% and 100%. Cooling is a way over radiated custom loop with a raystorm block and dual pumps with both GPUs and MB VRM in loop.

It has passed a 71 hour P95 Blend test this is just a quick Prime load to show temps:

[tcclaviger]

3 dead cpus out of the countless sandybridge owners who have read this thread? And all 3 at "safe" voltages that the vast majority of users on here run at if not higher for the last year or so?

Vcore was not the root cause for these cpus degrading.

Other factors must have been involved to cause these 3 cases of cpu degradation. Also note there are only 3 reports of cpu degradation on the forum out of many hundreds if not thousands of sandybridge overclockers viewing or posting here. There are too many people running 1.5v or more still having no issues at all for vcore to be an issue. Keep your temps under control and you're good.

I have seen it referenced that VTT voltages over 1.1 (For SB-E) lead to premature degradation rapidly. Perhaps SB is similar in that, it is not Vcore but another voltage, just a thought.

[PiLsY]

I really wouldnt like to hazard a guess at what causes it. Its clear that some effect realised during overclocking can cause degradation, but the positive result pool is so low compared to those of us who have no issue at the high end of the voltage scales that whatever is causing it affects only tiny minority of users.

Of course this doesnt mean it sucks any less if it happens to you, but I'd have a clear conscience returning a degraded cpu given that we dont know the cause.

I ran 1.52vcore, 1.15vccio and 0.975vccsa @ 5ghz for 13 months. First 8 months intensive memory testing using long runs of prime, after that it got moved to my gaming rig. Only replaced it as I got a really nice second hand 2500k for cheap, its not im my server running 4.2ghz 24/7 at default dvid.

[Blackbird_CaD_]

One guy on the internet said that if you immediately overclock (raise the vcore) after you have bought the CPU it will degrade MUCH faster, that you will then have to increase the vcore after one or a couple of months.
And that you should run the Sandy Bridge CPU with stock voltage for a couple of months before you increase the vcore.

I don't know if this is true.
What do you guys think?

[Hondacity]

One guy on the internet said ...

lmao

[Blackbird_CaD_]

lmao

Yes, he was just like you guys, he knew about overclocking and stuff.
Would it have been different if someone in this thread would had said it instead?

[[XC] Oj101]

No, because there is no logic to that statement.

[herderien]

after 3 month@1,40vCore (1,45 load) and 5Ghz, i experiences a blue screens while playing diablo.
I thought it was my case cooling that wasn't enough since i went for H80 instead of air cooling.
First i went down to 4,8 and 1,36vCore (1,41 load) and experiences no issues.
I then changed my case for a Corsair 500R and added 2 120mm fans on the top, and one in the bottom.
Temps went drasticly down on mobo, pcie and such, i had no issues, so i puted my 5ghz settings back and got another bluescreen.

Proc was prime stable, diablo is not that cpu dependant, and coretemps went barely to 50°C...so i assume is has degraded.

[Demo]

Is it still prime stable now? If it is then your having low load stability issues.. D3 barely hits 15% CPU usage.

[herderien]

it's a least 5 mins prime stable, wich is what i consider stable for my use (i know i should do longer but it's boring me ^^)
something i noticed, is i think i just changed the vcore and multi (from 1,41 to 1,36) but the vcore seems more stable, @5 and 1,41 it would stabilize @1,45
i guess i missed something somwhere...

i agree for diablo and cpu load...maybe it was something else...but i experienced no issues as of now, i'll check and keep you posted

[xarot]

5 mins Prime stable? Wait, what did I just read?

It seems stability is really subjective...with a test that sort it might not even pass 3D06...

[herderien]

Well i'm not a 4Cores 8 threads CPU intensive user, so i stress test a little, if i encounter issues afterwards i add more vcpu or lower my OC.

my way of "stress" testing usualy fits my needs, wich is why i trust it. System was stable for my use during 2/3 month of gaming (SWTOR,SC2, Diablo when it came out).

I had my first BSOD playing diablo after a week or so, i thought it was my watercooling causing lower airflow then my previous air solution on side parts (video card, MB...), so i lowered My OC, changed my case for a cooler one (ie with more fans).
I then rolled back to my 5Ghz OC, and got a BSOD shortly after, playing diablo (wich is not that CPU intensive - around 60/80% on 1 or 2 cores depending - but needed the 1,45vcore).

maybe it was not the load but just the 1,45v...

[rge]

3 dead cpus out of the countless sandybridge owners who have read this thread? And all 3 at "safe" voltages that the vast majority of users on here run at if not higher for the last year or so?

Vcore was not the root cause for these cpus degrading.

Other factors must have been involved to cause these 3 cases of cpu degradation. Also note there are only 3 reports of cpu degradation on the forum out of many hundreds if not thousands of sandybridge overclockers viewing or posting here. There are too many people running 1.5v or more still having no issues at all for vcore to be an issue. Keep your temps under control and you're good.

+1
The only mean time to failure curves I have seen in a white paper was long ago on a pentium cpu, where it was expected that 2-3 cpus per 10,000 would fail/die within 3 year warranty (and that was target, but not always achieved), and up to 10 years progressively more, and that is at stock settings. There are always reports of people's cpus that severely degrade to point of instability/die running at stock settings, just rare on cpus.

Even if 3 out of all comers had issues, to me that would be not unexpected, and just 3 lemon cpus.

[Ket]

If it helps put anyones mind at ease I've run my SB CPU @ 4.5GHz 1.33v running BOINC since the day SB was released and no harm has come to it. Take it for what you will. Those who report they killed their CPU I would think at the very least simply had bad luck and got a chip with weak core(s).

[Demo]

Just reporting in, mine is still running flawlessly and its been more than a year.

[AMD LOving man]

Bump the topic so more people share experience. I am mostly interested in max settings for 24/7 BOINC (100%) load?
I saw only two examples. Any more?

P.S. Please do not dilute the topic with bragging and help advices.

[nuttah]

2600K 4.8ghz + 1.45 rock stable. Got it a few days ago, OCed after the first day. Im gonna try lowering the voltage. Btw I had a q6600@3600mhz for 5 years, rock stable... always on watercooling. Ill post more info about the 2600K in a few days

EDIT: So far stable for an hour (using intelburn test on high) at 1.35v, 4.8Ghz, idle temps are 27-35C, having low speed on 2x12cm fans. So temps can get lower. Pll is on auto (almost everything on auto except for cpu core), idle shows 1.8v pll. Ill post core and pll numbers and cpu temp during stress proccess, I just set up the lcd screen with those numbers.

On load 1.36v 4800mhz stable, 47C max, pll 1.8. Stressing doesnt seem to change anything voltage related...Gonna keep this 24/7. I might try 1.4-1.45v and see how high it can go...

[AMD LOving man]

Let me share my experiment:
November 2012 - Intel ES 3.1GHz, HT, 6MB L2 (it is neither 2500k, nor 2600k) From day 1 it is 4700@1.42V. I was trying to run it at 1.41V for a few weeks but I was getting random BSODs, so now 1.42V is OK for a week.
Temperatures - Prime ~80C. Under normal use (24/7 BOINC ) 71C.
Generally I am keeping it under water but it is still quiet hot, so this should be a good example of degrading . Let see how long it will survive.

[3lfk1ng]

Not a dead chip but figured I should contribute to the thread. I'm no stranger to degredation as my e8400 (x3110) did the same thing after ~16months.

i7-2600k
For the first 8 months, I could do 5GHz @ 1.45v. For the last 4 months, I could do 5.1 @ 1.55v and 5.0 @ 1.52v but no longer 5.0GHz capable @ 1.45v. I figured I had hit the threshold multiplier of my i7-2500k and my motherboard wouldn't allow me to push voltage above 1.55v.

Finally last weekend my computer decided to shut off in the middle of a Rune lan party. I knew right away that temps weren't the culprit as my processor never goes above ~50c even during OCCT. I rebooted to discover that my computer would fail every attempt to get into windows, even at 4.8 and even at 4.6GHz. If I mess around enough I can get into windows at 4.4GHz but it doesn't seem stable so for the time being I am stuck at a measly 4.2GHz (FWP).

So... 12months and -800MHz later, here I am.

[SIOUX]

Have a 2500K CPU running at 4500mhz 1.35v for 1 year and 7 months. No degredation what so ever.

[LagunaX]

Yeah I have a 2600k (1.328v) and 2500k (1.344v) at 4.8ghz no degradation either for over a year each at least too.

[PiLsY]

My 2500k is still at 5.0 / 1.51v, it got moved back to my gaming pc as my 3770k sucks. Must be 2 years old now.