OC used to be stable, but now isnt?

[Smoblikat]

So i flashed my mobo BIOS and an overclock that was stable before is now unable to boot into windows . I had 5.3ghz before, and now as soon as windows tries to start i get error code 9C. I think im using the exact settings so i dont know what is wrong . yes ive tried reverting to the older BIOS and this hasnt helped either.

 

[Idontcare]

Transistors degrade over time, as do the wires that connect them.

Increasing the operating temps, operating voltage, clockspeed (think rpm's on an engine block), will hasten the rate of degradation and decrease the chip's lifetime.

This is part of the reason why a stock-clock CPU seems so conservative in the clocks...the manufacturer has to ensure the clockspeed is sustainable for years and years, not just for 3 months.

We OC'ers come along and jack up our CPU's, basically burning the proverbial candle at both ends, its twice as bright but burns out much quicker.

Your chip has more than likely suffered irreversible degredation and simply cannot stably sustain the 5.3GHz clockspeed you had it jacked up to. This is the downside, the peril, of overclocking.

 

[Smoblikat]

Transistors degrade over time, as do the wires that connect them.

Increasing the operating temps, operating voltage, clockspeed (think rpm's on an engine block), will hasten the rate of degradation and decrease the chip's lifetime.

This is part of the reason why a stock-clock CPU seems so conservative in the clocks...the manufacturer has to ensure the clockspeed is sustainable for years and years, not just for 3 months.



We OC'ers come along and jack up our CPU's, basically burning the proverbial candle at both ends, its twice as bright but burns out much quicker.

Your chip has more than likely suffered irreversible degredation and simply cannot stably sustain the 5.3GHz clockspeed you had it jacked up to. This is the downside, the peril, of overclocking.

I dont think it degraded. I only ran it at 1.37v, and for several weeks at most. Im thinking its just some dumb feature that i forgot to set. Should i enable C3 and C6? And what about the max instantaneous watts allowed through the 1A rail or whatever?

It ran fine before the BIOS flash, and now after it doesnt run at all. So im positive its teh mobo.

 

[notty22]

Did you ever verify you were running 5.3ghz with a cpu-z screenshot ? Or a super pi run , that proved that speed was actually running. 5.3 @ the voltage you are talking about is not normal results.
I like the above Tyrell reference

 

[Smoblikat]

Did you ever verify you were running 5.3ghz with a cpu-z screenshot ? Or a super pi run , that proved that speed was actually running. 5.3 @ the voltage you are talking about is not normal results.
I like the above Tyrell reference

Yes i ran CPU-z, im not a noob . I also have my superpi results when i was at 5.3. I know i was at 5.3, and my Vcore was like 1.42 in BIOS and 1.37 under load.

 

[Idontcare]

I dont think it degraded. I only ran it at 1.37v, and for several weeks at most. Im thinking its just some dumb feature that i forgot to set. Should i enable C3 and C6? And what about the max instantaneous watts allowed through the 1A rail or whatever?

It ran fine before the BIOS flash, and now after it doesnt run at all. So im positive its teh mobo.

5.3GHz with only 1.37V? Sorry, I can't take that for serious.

I don't think you are trying to troll us, which leaves the only rational conclusion here being that you were somehow confused as to either the voltage or clockspeed involved.

Error code 9c is usually taken to mean the CPU is not getting enough voltage.

Did you manually set the voltage to 1.37V? Or did you just leave your mobo to do its "auto CPU volt" thing. Because unless you manually told it to do 1.37V, your mobo will dial up the voltage, all the way to 1.50V, as the SB's VID itself increases by auto-detection of the multiplier.

At any rate, given your current conundrum and the seeming outlandish claims of how your prior setup was operating, it all points towards the most likely scenario here being that you were confused as to how and what your original OC was setup to be.

edit:

Yes i ran CPU-z, im not a noob . I also have my superpi results when i was at 5.3. I know i was at 5.3, and my Vcore was like 1.42 in BIOS and 1.37 under load.

Hmmm...so not LinX stable then I take it?

I can do 5.3GHz with my 2600K but it doesn't pass LinX testing. SPi is fine though. I would not claim it as an OC, not here at AT anyways. We tend to define OC as 24/7 stable with LinX or Prime95, not SPi.

 

[soccerballtux]

pics or ban as they used to say
and by ban it meant "it didn't happen"

 

[jhu]

Blame quantum mechanics.

 

[Smoblikat]

5.3GHz with only 1.37V? Sorry, I can't take that for serious.

I don't think you are trying to troll us, which leaves the only rational conclusion here being that you were somehow confused as to either the voltage or clockspeed involved.

Error code 9c is usually taken to mean the CPU is not getting enough voltage.

Did you manually set the voltage to 1.37V? Or did you just leave your mobo to do its "auto CPU volt" thing. Because unless you manually told it to do 1.37V, your mobo will dial up the voltage, all the way to 1.50V, as the SB's VID itself increases by auto-detection of the multiplier.

At any rate, given your current conundrum and the seeming outlandish claims of how your prior setup was operating, it all points towards the most likely scenario here being that you were confused as to how and what your original OC was setup to be.

edit:

Hmmm...so not LinX stable then I take it?

I can do 5.3GHz with my 2600K but it doesn't pass LinX testing. SPi is fine though. I would not claim it as an OC, not here at AT anyways. We tend to define OC as 24/7 stable with LinX or Prime95, not SPi.

My voltages were offset mode @ + 1.105v, so whaever that is is what my OC was. No i dont even know what linx is , it was prime stable for about an hour or so, then i got a BSOD. But it was game stable.

Im very serious if i had known that 5.3@ 1.37 was so good i would have taken a screenshot of it, my CPU does 3.3ghz @ .960v fully prime stable for ever. Is that good?

 

[toyota]

I5 2500k @ 5.3ghz, 28C idle, 50Cish max load. My cooler is awesome.

nobody was buying this comment of yours either btw.

 

[Idontcare]

My voltages were offset mode @ + 1.105v, so whaever that is is what my OC was. No i dont even know what linx is , it was prime stable for about an hour or so, then i got a BSOD. But it was game stable.

Im very serious if i had known that 5.3@ 1.37 was so good i would have taken a screenshot of it, my CPU does 3.3ghz @ .960v fully prime stable for ever. Is that good?

Try more Vcc to see if the error code 9c goes away.

You did remember to set the "PLL Overclock" to "enable", right?

 

[Smoblikat]

Try more Vcc to see if the error code 9c goes away.

You did remember to set the "PLL Overclock" to "enable", right?

Yes i did, my PLL voltage is 1.75, is this a good voltaeg for OCing?

nobody was buying this comment of yours either btw.

Well, with C1E and speedstep enabled my idle volts are 1600mhz at under 1v.............so why is that temp so incredible? For load i just have low ambients since in in a basement in the middle of winter. If i get my overclock back i swear i will proove you all wrong. I just dont know what happened............i went from v1.1 to v1.2 and i couldnt OC at all, now im back to 1.1 and i can get my stock undervolt clocks back.

 

[Schmide]

Although superpi may be good at finding problems with an OC (cache) it is far from worthy for declaring an OC stable. Unless you pass OCCT, LinX or prime it's a fluke.

 

[Smoblikat]

Although superpi may be good at finding problems with an OC (cache) it is far from worthy for declaring an OC stable. Unless you pass OCCT, LinX or prime it's a fluke.

To me an OC is stable when it can run crysis. Since i dont play P95 or linx or OCCT all the time, i play games.........so if it can pass my stressiest game it can pass them all. And after a reflash of my BIOS im almost back to my origonal clocks n volts. Under full P95 load im at 1.37 but it fails after a few minutes. Normal operations put it between 1.42 and 1.44 for 5.3ghz, pics in a sec.

 

[Smoblikat]

 

[ed29a]

50Cish under load?

 

[Schmide]

Try and take out the cmos battery. It clears the cmos better than the jumper.

Regardless - you should test your OC for long term usage. 5.3 is a bit much for long term usage. Just do 4.5 and accept you're not a hard core OCer...or go out and buy some watter and do it right.

Edit: Run superpi and look at the CPU usage. It rarely goes above 60% on one core.

 

[Smoblikat]

50Cish under load?

Thats P95, that quick run of superpi brought it to a whopping 25C

 

[Smoblikat]

Try and take out the cmos battery. It clears the cmos better than the jumper.

Regardless - you should test your OC for long term usage. 5.3 is a bit much for long term usage. Just do 4.5 and accept you're not a hard core OCer...or go out and buy some watter and do it right.

Edit: Run superpi and look at the CPU usage. It rarely goes above 60% on one core.

Yup, my CPU is barely being stressed. I didnt know that superpi didnt use all the CPU.........i wish i could get my old setup back. Ill try removing CMOS tomorow, its like 11:00 right now and ima crank some tunes then hit the sack.

 

[ed29a]

Thats P95, that quick run of superpi brought it to a whopping 25C

Liquid helium cooling? I have a very hard time believing that a stressed 5.3 Ghz overcloked 2500K on an EVGA SuperDuperMega Cooler only hits 50Cish per your previous post.

 

[Idontcare]

Yes i did, my PLL voltage is 1.75, is this a good voltaeg for OCing?

That's not the same thing. Don't confuse "overclocking your PLL" with the separate BIOS feature called "PLL Overvolt". They are two separate things.

PLL Overvolt is a binary thing, its either enabled or disabled. Typically it needs to be enabled in order to get a SB to OC above 4.8GHz.

To me an OC is stable when it can run crysis. Since i dont play P95 or linx or OCCT all the time, i play games.........so if it can pass my stressiest game it can pass them all. And after a reflash of my BIOS im almost back to my origonal clocks n volts. Under full P95 load im at 1.37 but it fails after a few minutes. Normal operations put it between 1.42 and 1.44 for 5.3ghz, pics in a sec.

That's not what stress testing is for. Stability is a statistical thing, not a binary thing. There is no such thing as "Stable". There is only "stable for X time" where X is a number derived from statistical analysis of the fail rate.

To arrive at some level of understanding your fail rate (likelihood of suffering an instability) you need to stress test your system. We use programs like LinX and Prime95 to accelerate the stress testing.

Maybe not everyone realizes this is why they are using it, but that is the reason.

The science behind accelerated stress testing is designed to flesh out the stability profile for your setup so you have some degree of confidence and knowledge (versus just winging it on a hope and a prayer) regarding the likelihood of your system crapping out on you in 1 day, 1 week, 1 month, etc.

When you see someone reference that their system is 12hrs LinX stable the reason this is done is because 12hrs LinX stable means the likelihood of the system being unstable with a regular workload (less stressful than LinX) is lower by roughly an order of magnitude (120hrs or more).

It doesn't mean it is "stable", just means it is acceptably stable.

If you don't stress test your OC then you are leaving yourself open to all forms of data corruption and nasty surprises over time.

 

[Avalon]

It does sound like it's degraded a bit, albeit a bit quicker than I would have imagined. Regardless, a BSOD in an hour of P95 is pretty scary. You may be fine with it just being able to run Crysis, but if your CPU is producing errors and you're doing anything on your system that's important, you could be silently corrupting some of your data.

I'd really recommend backing down some and getting 8hr+ P95 or LinX stable.

 

[PsiStar]

I don't you describe your cooling system ... having just glanced up thru the thread, but I suspect it has degraded or become less efficient in some way.

 

[Smoblikat]

Hmmmmmm...........the general consensus is that my chip has degraded. Would it do that with low voltage such as this? (max of 1.43)

 

[Rvenger]

@ 5.3ghz I bet it was never stable. How long have you had this cpu, 2 months? Probably just the initial break in. My 2600k was perfectly stable @ 4.4ghz with 1.32v. A month or so later it wanted more voltage. No biggie. Just apply more volts or scale back on that massive OC. I wouldn''t worry about it. I owned 3 of these chips and they behaved in a similar way you are explaining.