X99 Issues - Stability

[StealthArsenal]

Hi everyone,

 

Looking for a little advice on an issue that originally started this past weekend, but seemingly reared its ugly head yet again last night.  Just for additional information, this system was build just about a year ago and has spent most of its life thus far at 4.5ghz with a Swiftech H-240x cooling it.  Over this past weekend I started experiencing Core instability on my 5820k.  I was noticing that a couple of the cores were unusually hot compared to the others at idle and the system was locking up and freezing.  We are talking idle temps approaching the upper 50's C on the troublesome cores while the remaining cores were running at 38C max.  I chocked that up to failed thermal paste and replaced it.  At that point, temperatures seemed to normalize (35C-41C) and all was fine.  Last evening, I again started to experiencing stability issues but this time the temperatures all looked good (again 35C-41C) and I received a BSOD CLOCK_WATCHDOG_TIMEOUT.  I rebooted the machine once or twice with the same BSOD showing upon entering the Windows environment.  I shut it down once more to get into bios (fat fingering the wrong function key) and upon boot Stop Code 0XC000000e.  Disconnected all my drives but the boot device and was able to re-enter Windows. With that said, I entered bios and dropped by overclock to 4.4 and manually reset the cpu voltage at 1.28 volts in order to obtain stability or at least what I think is stable at this time.  

 

With all that out there now, here are my questions:

 

1. Am I seeing a degradation of the 5820K (i.e. a premature failure on the cpu).

2. What would seemingly cause the system to suddenly go from stable to unstable.

3. The chip, I should say was a stable chip at 4.7 GHz, but I never ran it that high for more than a few hours back when I built the system.  Does that in any way factor into the chip life.  In other words I don't mean to say my short period at 4.7 hurt the chip, but was this chip maybe a problem from the start?

 

Thanks for the input.

[mprodani]

What cooler do you have?

If its a closed loop it might be that the fluid inside might be bad...

The BSOD is for clock instability on a single core.

Try clearing your Bios and retrying to do the overclock and make sure you have set all your cores to work at the same speed.

EDIT--Saw you have a swiftech cooler. It probably is a cooler problem and those temps at idle for a closed loop seem just way too much even 

with an OC

[DXMember]

just roll with it for now, if the issue reappers we'll talk again later

[mprodani]

Also remember that 4.7ghz is a lot for that CPU and will show degradation overtime

[xgn]

Did you get it on windows 10 ? because i got this 2 times on windows 10 while my OC were stable on windows 7 .

I would ignore that message with waiting for ram timeout. If you get actual bsod try increase vcore.

[runit3]

1. Am I seeing a degradation of the 5820K (i.e. a premature failure on the cpu).

2. What would seemingly cause the system to suddenly go from stable to unstable.

3. The chip, I should say was a stable chip at 4.7 GHz, but I never ran it that high for more than a few hours back when I built the system.  Does that in any way factor into the chip life.  In other words I don't mean to say my short period at 4.7 hurt the chip, but was this chip maybe a problem from the start?

 

Thanks for the input.

With the info in #3 it sounds like it was never stable at 4.7 to begin with (extremely likely, you would need a miracle chip to hit 4.7 at a voltage under 1.4v). It takes more than a few hours just to stress test a system with ONE test, let alone all the tests  necessary to call it a stable OC.

 

Running your chip at 4.7 didn't do any harm to it unless you had your voltage somewhere up around 1.45-1.5v, then voltage degradation certainly could have happened (only a little). Clock speed does not kill chips, voltage kills chips.

 

My best guess at what actually happened was that you were finally doing something with the computer to trigger instability, could have been as simple as watching Netflix.

[StealthArsenal]

I had the Swiftech cooler apart a couple of months ago to clean it out, add new fluid and install a kill coil.  Nothing has seemed out of the norm with it at this point in time.  I know if I drop back to stock clocks, the machine idles at 22C.  I know these are not load temperatures, but I didn't load it up at stock clocks.  These BSOD's are on Win 10 and it was the first time I have experienced them.  The BSOD last evening was while in Diablo 3.  Thought it might be a good time to check out a patch and after about an hour the crash occurred.  This past weekend, the machine was idling when everything went crazy.  Now from what I have read, my idle temps at 1.28 to 1.3 volts seem relatively consistent at 4.4-4.5 GHz.  I am going to run the machine at the reset cpu voltage and see if it happens again.  I also want to reconfirm the temperatures later this evening with the increased voltage.

 

The 4.7 I ran about a year ago was stressed in P95 for 24 hours at roughly 1.42 volts before I dropped it down to 4.5. I don't consider 24 hours a true stability, but it was something.  I won't harp on that number though as I don't run the CPU at that speed so its void at this point from what I have read here.