Anyone experienced in cpu's degrading? Not sure if my 2500K is..

[ZaO]

I bought my 2500K new, early in 2012. I've been running it at 4.5Ghz (no higher) the entire time I've had it (bar a couple months or so over all). When I first OC'd it I was needing about 1.36v (give or a take a little due to fluctuations in vcore). I upgraded my bios last year and found it needed a bit more to remain stable. So was running it at around 1.38v give or take. Earlier this year I found I was getting blue screens (101) about once a month or so. So I upped the vcore by the smallest increment possible (0.005 I think?). It has been running fine all summer. But then Splinter Cell Blacklist caused me a blue screen the other day, right after I installed the 13.10 Beta driver for my 7970 might I add. I assumed it was the Beta driver as it came up in Bluescreen view. But in line after it I'm seeing ntoskernal as well which is the one I'd been seeing before this. So I went back to the 13.8 Beta driver. Things seemed alright for a few days. Then my computer blue screened at some point lastnight while it was running a virus scan with September's Windows malicious software removal tool (another 101 with just ntoskernal showing). So today I have upped the vcore by one small increment again. vcore is anywhere from 1.39-1.414v under full load. Max temp with p95 is 72c. So I'm hitting the limit of what I feel comfortable with here on this chip. I am considering taking it down to 4.4Ghz if this continues. I know that vcore might not necessarily be the issue (even though Blue screen - 101 suggests it). But I can figure that out later.. Not in the mood for playing about too much right now..

So what I'm wondering is - do you think my chip might be degrading? I've never had this happen to me before because this is the first chip I've really overclocked. Are these tell tale signs in your experience? I'm wondering if I should ease off the oc so it can last a bit longer. Don't want it dying just yet. Hoping to get a couple more years out of it! Thanks all!

Edit: Btw - I've never volted it higher than what it's at now.

[Jonj1611]

I have had it happen with a Q6600, after a couple of years had to increase the voltage to keep the same speed, it does happen my man.

[mikeo01]

I will be honest, I haven't had experience with overclocking Intel chips thus I do not know their upper limits in terms of voltages.

However, usually newer chips are more durable and can withstand higher voltages for longer. Essentially a chip shouldn't degrade that quickly. However that said heat is a major killer; so heat can ultimately reduce the chips life span.

Also you have to take into consideration your motherboard. Your motherboard's VRMs have been supplying an increased demand in power, thus higher current leads to an increase stress on the mosfets.

Although Intel chips are decent in terms of power consumption there are many factors at play. Weakened VRMs, heat has caused slight degradation on either the mosfets or the chip itself. Also depending on how you overclock depends on how you effect your hardware. Pure multiplier overclocking is the norm, and obviously targets the northbridge, VRMs and chip's internal NB.

Ntoskernal is in charge of processes and memory management. So you could take a look at your RAM if you've overclocked it.


Edit:
However from posts I have seen (not valid source of info!) isn't 1.4v+ on Intel chips high? Not sure about that.

[Jonj1611]

Max Vcore on that chip is 1.52V though not many need to go that high.

[Bonebreaker777]

Somewhere I read that 1.3-1.35 vCore for the CPU (not sure if that was for Sandy Bridge as well, but for sure it was for one of the Core generations) and 1.65V for the RAM is suggested long term for Intel chips. Judging by Za0's System description I don't think that he is overclocking his RMA.
My money would be on the MoBo. I would personally how high I can go with 1.35V and consider after if the decreasing life is worth it as opposite the higher performance, with a bit of luck minimal it may be.

[mikolaj87]

As others mentioned it is possible that your mobo may cause it but also power supply can degrade. For example my cooler master provided almost flat 1.35v to my cpu. Now voltage on my cpu vary from 1.32 to 1.37 while mobo still is set to the same voltage.

[Agent]

So what I'm wondering is - do you think my chip might be degrading? I've never had this happen to me before because this is the first chip I've really overclocked. Are these tell tale signs in your experience? I'm wondering if I should ease off the oc so it can last a bit longer. Don't want it dying just yet. Hoping to get a couple more years out of it! Thanks all!

It's possible. The term is called Electromigration - it happens faster when you overvolt / clock.
Sadly there is no way to know if this is the issue, motherboard power regulation issues or (somewhat less likely) the power supply. You'd just have to swap things out and play with the clocks.

Keep in mind that you've been running it at over 20% additional speed since new (by the sounds of it), along with a voltage bump. You don't know how good the silicon was to start with

[ZaO]

Thanks for all replies!

I used this guide to oc it: http://www.overclock.net/t/968053/of...nc-spreadsheet
It answers a lot of questions about voltages and temps based on a large community feedback and other sources. Here's a couple of quotes from it:

"No one is absolutely certain of what the safe vcore is for the sandybridge chips. What I can tell you is that many say the max safe vcore for these chips are 1.3 region, however, intel states that the max 'VID' is 1.52 and many say that around 1.4 if your on air and 1.45 should be the max if your running water cooling. Personally I will not go above 1.5v for 24/7 with these chip but that is totally upto you. The main thing to understand is that 'YOU' have to come up with the conclusion of what the max is. That way no one is blamed if the chip degrades (none reported so far, even with so called 'high 1.4+ vcore')"

"Regarding temps, CPU throttles at 95c, some say keeping it below 85c is good, some say keeping it below 80 is better, other's say below 75c is really good and there are quite a few that say 70c should be the max. Which ever one your comfortable with and if you have substantial cooling, YOU DECIDE YOUR MAX, just remember it throttles at 95c. If for example you hit 85c in stress testing then in everday usage it shouldn't be higher than 75c which I think is fine, I personally like to keep mine below 70/75c during general everyday usage"

Intel says 72c for these chips on their site btw. As I just hit 72c with my last test, I started to get concerned a little as I don't want to push anymore..

I haven't overclocked the ram and 72c is the highest the cpu's been yet. I put a submission in on that club with a 99% load on ram and 18.5 hours on prime95 blend to pass all fft's. That was of course when I hadn't had the chip as long. I have ran some p95 blends since then. With 99% ram usage but running each fft for 10 minutes so I can pass them all in under 13 hours. I actually passed one after the last time I upped the vcore (the time before yesterday I mean). But you can get bsod's outside of testing even when you've passed a p95 test as I have learned.. I don't think my psu is bad! I bought it new around the same time and I doubt it's been much more than around 50% under load. My setup now is the most demanding this psu has seen. Though I guess I couldn't completely rule it out. I bought the mobo new as well and it's only seen this setup (other than having a gtx670 and also a different 7970 before that). Maybe it is the mobo. Who knows lol :/

I guess I will see if it does alright at it's current settings. If not I shall have to take it down a notch and do all the proper testing from scratch to get the voltage down as low possible with 4.4Ghz. But if anyone has anything to add then please feel free! Thanks for the advice