You should be able to drop it even lower. I run mine @ 4.8GHZ running at 1.336 volts.

Ben, no way. If I drop it to 1.37V, it BSOD's after 15 minutes of Prime95 with the 0x124 error. I need at least 1.38V, perhaps even 1.39V. I've heard that the later i5 2500K's seem to be much more voltage "hungry" than the earlier i5 2500K's, I dunno why.

Ah ######. Running 4.7 GHz 1.38V, got BSOD after 8 hours of Prime95. What voltage should I try now? 1.385, or 1.39?

Ben, you are running a 2700K,

I am?!?

Ah ######. Running 4.7 GHz 1.38V, got BSOD after 8 hours of Prime95. What voltage should I try now? 1.385, or 1.39?

I never quite understood the OC'ers obsession with running P95 for endless hours to prove stability!Is your PC going to be primarily used for running P95, or for playing FSX and other games?For my overclocks, if the temperatures and voltages are within safe operational limits, I can successfully run an IntelBurnTest set at max settings and FSX and all my other games are stable during the hours that I play them, then my overclock is successful and I put it into operation!Surely stability should be measured by the PC being "stable for purpose"?...I am interested that you reckon you get a noticeable improvement in performance from 4.5 to 4.7 - I have been running my 2500k @ 4.5 (Turbo boost) for a while now and haven't bothered to push it any further, but if people are getting an improvement, I may push it a bit more!FWIW I have my multiplier set to 45, all the voltage, PLL and related settings to "auto" and the core voltage set to negative offset of 0.010 in order to bring the max usable voltage down slightly. In heavy usage, it peaks at around 1.312 and everything runs fine.Cheers

I never quite understood the OC'ers obsession with running P95 for endless hours to prove stability!

If you didn't understand it, then you don't know how to properly overclock.

Surely stability should be measured by the PC being "stable for purpose"?

Wrong.

 

If you didn't understand it, then you don't know how to properly overclock.

I have never had any issues with any of my previous OC'd machines - they have all been fast, cool and completely stable. I disagree that because I don't follow an excessively dogmatic methodology that I "don't know how to properly overclock" :Kiss:

Wrong.

Enlightening, thank you? :Thinking:To add; I have also found Prime95 lacking as a "gold standard" on more than a few occasions - other tests have pushed systems harder and exposed failures, voltage issues and unacceptable temperature spikes far more quickly...

FWIW I have my multiplier set to 45, all the voltage, PLL and related settings to "auto" and the core voltage set to negative offset of 0.010 in order to bring the max usable voltage down slightly. In heavy usage, it peaks at around 1.312 and everything runs fine.

If you can run 4.5 GHz you've gotten yourselves a golden chip. I need 1.34-1.35 for 4.5 GHz, and 1.39V for 4.7 GHz. I never get the good hardware. :(

If you can run 4.5 GHz you've gotten yourselves a golden chip. I need 1.34-1.35 for 4.5 GHz, and 1.39V for 4.7 GHz. I never get the good hardware.

The variance in stable OC'd voltages within a given model of CPU can be pretty frustrating! I have got lucky with the last few I have had it seems...Who knows - if I spend the rest of my life running Prime95, I may expose issues, but who wants to do that when I could be flying! :Just Kidding:Suffice it to say that it runs at a nice, safe voltage, it is solid as a rock with every game and app I can throw at it, whether that is FSX or 8 straight hours of BF3 and I have never managed to push the temps above 57 degrees in benching ( an hour or so of P95 and a few runs of max-strength Intel Burn Test). In real world usage, it never climbs out of the late 40s.Cheers and good luck with the tweaking!

The variance in stable OC'd voltages within a given model of CPU can be pretty frustrating! I have got lucky with the last few I have had it seems...Who knows - if I spend the rest of my life running Prime95, I may expose issues, but who wants to do that when I could be flying! :Just Kidding:Suffice it to say that it runs at a nice, safe voltage, it is solid as a rock with every game and app I can throw at it, whether that is FSX or 8 straight hours of BF3 and I have never managed to push the temps above 57 degrees in benching ( an hour or so of P95 and a few runs of max-strength Intel Burn Test). In real world usage, it never climbs out of the late 40s.Cheers and good luck with the tweaking!

Yeah that's true.But an overclock definately needs to be stable. An unstable overclock is not gonna damage your hardware or something, but you can get many ntdl.dll and API.dll CTDs in FSX if you have an unstable overclock. So far I managed to reach 8 hours Prime95 with 1.38V. Now I upped to 1.39, and if I reach 12 hours, then I know it's rock solid and then I'm also sure that I will never (or very rarely) have any CTDs

Yeah that's true.But an overclock definately needs to be stable. An unstable overclock is not gonna damage your hardware or something, but you can get many ntdl.dll and API.dll CTDs in FSX if you have an unstable overclock. So far I managed to reach 8 hours Prime95 with 1.38V. Now I upped to 1.39, and if I reach 12 hours, then I know it's rock solid and then I'm also sure that I will never (or very rarely) have any CTDs

I completely agree that it has to be stable and within safe operating limits...I am just saying that for me, if everything I do is completely stable and the temps/voltages/performance are all good, then I feel I have nothing to be gained by running P95 for 12-24 hours over the tests I have already run!Out of interest, have you tried the IntelBurnTest? I find at max stress settings it pushes the CPU and RAM harder than P95 and can expose potential issues faster, i.e. 5-10 minutes or so per run...Perhaps that first, then try running P95 overnight if you want to double check. If it can't get through the Burn Test then it isn't likely to get through the P95 either and you will have found out more quickly.

Enlightening, thank you?

No, wasn't my intention. There are virtually hundreds of threads, tutorials, forums and the whole internet at your disposal to check why.

I have also found Prime95 lacking as a "gold standard" on more than a few occasions

There is not such thing as a gold standard. P95 is just one of the tools. You need couple of, to really confirm the stability. And even then, you might run into problems.

No, wasn't my intention. There are virtually hundreds of threads, tutorials, forums and the whole internet at your disposal to check why.

Indeed there are and you will find more than a few which say the same thing as me!

There is not such thing as a gold standard. P95 is just one of the tools. You need couple of, to really confirm the stability. And even then, you might run into problems.

Completely agree here.Again I say though;If P95 works for an hour or so,If Intel Burn Test completes several runs at max stress settings successfully,If Furmark, Folding and 3D Mark all run fineIf voltages are well within safe limitsIf temperatures are all within safe limitsIf I can game all day with no hitches, glitches, crashes, BSODs or other problemsIf in everyday usage my PC is stable...Then as far as I m concerned, I have a stable and successful over-clock. At this point a 12 to 24 hour run of P95 is not going to tell me anything of interest and I will have spent those hours doing something infinitely more fun and productive with my computer!Cheers

Indeed there are and you will find more than a few which say the same thing as me!

Seriously? I am a more than occasional visitor on Hardforum, XtremeSystems, Forum de Luxx, Overclockers UK, and I never came across the similar statement.But OK... let's just leave this here.

Seriously? I am a more than occasional visitor on Hardforum, XtremeSystems, Forum de Luxx, Overclockers UK, and I never came across the similar statement.But OK... let's just leave this here.

*shrug*If you like!It is not my intention to be argumentative or cause friction for the sake of it and I can only post from personal experience - but I always enjoy discussion and learning...We obviously have differing opinions and methodologies used to attain our goals, but at the end of the day, we both seem to achieve the same thing no?If you disagree from a technical standpoint on my comment below and from the rest of this thread, would be glad to listen, to understand and to discuss - it is what a forum is for after all!...equally, if you can't be bothered then thats fine with me too.Cheers:)

If P95 works for an hour or so,If Intel Burn Test completes several runs at max stress settings successfully,If Furmark, Folding and 3D Mark all run fineIf voltages are well within safe limitsIf temperatures are all within safe limitsIf I can game all day with no hitches, glitches, crashes, BSODs or other problemsIf in everyday usage my PC is stable...Then as far as I m concerned, I have a stable and successful over-clock. At this point a 12 to 24 hour run of P95 is not going to tell me anything of interest and I will have spent those hours doing something infinitely more fun and productive with my computer!

To add an analogy:If I built a car from a kit, I would read the instructions thoroughly. I would research specifications and tolerances. I would seek advice from like minded people.After building it, I would check for structural and mechanical integrity and then I would take it out on the road, on the track, test the handling and test the performance in a variety of real world situations and have fun doing so.If it passed these tests and the car performed satisfactorily, safely and proved reliable, then I would be happy to drive the car from then on.At this point, I wouldn't see much benefit in then jacking the back wheels up on blocks, sticking it in gear, jamming a brick on the accelerator pedal, leaving it overnight in the garage and then declaring the car unroadworthy if I came back in the morning and the engine had exploded...