i7-2600K Overclocking Results

D3lirious · 17 Jan 2011 at 22:37

I put the following together on Saturday, and have had a go at overclocking.

Asus P8P67 Deluxe
i7-2600K
G.Skill RipjawsX 8GB PC3-12800C9 1.5v

The log below shows my overclocking steps. I've not gone any further than this yet due to not liking idle vcore nearing 1.4v. It looks like 4.6GHz is a relatively safe overclock, and I will probably keep it there at about 1.34/1.35v in BIOS until we get more feedback/data on safe voltages for these CPUs.

I may also play around with Load-line Calibration to see if I can equal out the idle/load CPU-Z vcore balance. It seems to be overvolting idle and undervolting load too much at present, the latter probably not helping stability.

I will do some overnight prime runs to better test stability - these quick runs were just to give me an indicator.

I've listed all BIOS settings too, although many of these are guesswork. I'm certainly open to tips for a higher, stable overclock.

Sayso · 17 Jan 2011 at 22:57

mmm you want to be priming for at least 4hours to test an overclock at the moment you have proved nothing.

For a quick check use IBT/Linx do a run of 20 at least if it passes this then do a prime blend test over night.

these CPUs seem a bit fragile and are great unoverclocked, i would wait for definite resuslts from people who need an epeen before clocking to a maximum. Go for the lowest volts/overclock that "you" need, lets face it not many people need a 2600K at 4.6Ghz its just a nice thing to write in there sig

D3lirious · 18 Jan 2011 at 09:16

Sayso said:
mmm you want to be priming for at least 4hours to test an overclock at the moment you have proved nothing.

For a quick check use IBT/Linx do a run of 20 at least if it passes this then do a prime blend test over night.

True, but I did mention the quick runs were only for an indicator. For example, I look back at the log and see that 1.35v BIOS and 4.6GHz might be good for a stable setting. I left it overnight at this and it was still OK this morning after 7.5 hours. If it's still fine after work I will stick to that until we know more about safe voltages.

Cheers for the tip on IBT/Linx, I will check this out.

Cupra · 18 Jan 2011 at 10:24

prime blend isnt the best overnight test when overclocking cpu's you should use small fft's for checking stability of a cpu

JonnyGLC · 18 Jan 2011 at 11:09

Cupra said:
prime blend isnt the best overnight test when overclocking cpu's you should use small fft's for checking stability of a cpu

Not with Sandy Bridge.

Systems can run all night and day with small fft's, then the second you throw blend at them they fail.

D3lirious · 18 Jan 2011 at 12:06

I've been using small ffts in all my prime runs. I was under the impression that that setting stressed the CPU the most, and that the other two settings put more stress on RAM.
My plan is usually to stress small ffts until CPU appears stable overnight, then start looking at best RAM frequencies/latencies which I test with large ffts (or whatever the second option is).

Cupra · 18 Jan 2011 at 12:12

D3lirious said:
I've been using small ffts in all my prime runs. I was under the impression that that setting stressed the CPU the most, and that the other two settings put more stress on RAM.
My plan is usually to stress small ffts until CPU appears stable overnight, then start looking at best RAM frequencies/latencies which I test with large ffts (or whatever the second option is).

yeah this is what i do