oc voltage for a i5 2500k? how low can you gooooo?

[philipma1957]

okay I got my open box ASUS maximus IV gene mobo last night. It seems to work fine and I am doing a 4.2 GHz oc of my i5 2500k . On auto the volatge is 1.31 volts too high . The Asus has a simple tweak that lets me lower the voltage. I have dropped it to 1.23 volts . I have a stable overclock I ran prime 95 over night and the max temp was 74c.

My question is how low have you (i5 2500k owners) been able to set voltage. I seem to remember a post from some one that they had a 1.19 or a 1.21 and it worked. Sooo if you have an i5 2500k or i7 2600k or even an i7 2700k hoiw long can you go and be stable. At 4.2 GHz. Thanks

 

[Kenmitch]

The 2700k I used to have did 4ghz @1.14v's but I stopped prime at 8hrs or so. LinX with AVX the loaded vcore was 1.15v's

I don't for the most part do alot of testing at such low overclocks. I just did the 2700k for the hell of it to see.

 

[aaksheytalwar]

There is a huge volts difference btw 4 and 4.2 and btw 4.2 and 4.4. It is like nearly 0.075-0.1V for every 200Mhz at this level
And beyond that it only gets worse.

 

[Kenmitch]

If you want to use prime for your testing and want to decrease your vcore at the same time to see how low it will go it's gonna take forever if your into the overnight runs to claim stability.

I'd suggest trying the following settings for quick runs. If you can make it for 20-30 min. or so then you could try dropping vcore again or just try the normal overnight settings. Not sure why or how it works but it does work and most of the time will bsod quickly if your votages are marginal.

You can also try setting the values to 1344k and test again if the values in the image are stable. If you like you can also bump up the memory value but you want to keep about 1.5gb's or so free for windows and some web browsing

There is a huge volts difference btw 4 and 4.2 and btw 4.2 and 4.4. It is like nearly 0.075-0.1V for every 200Mhz at this level
And beyond that it only gets worse.

YMMV as usual. Depends on the chip of course. Better chips will make the 200Mhz jump with as little as 0.020-0.030v's between jumps. Of course even better chips will require large vcore bumps once you start getting up there

 

[philipma1957]

I got the numbers down to 1.22 volts and it runs well.

at 1.20 volts it does not boot.


I wanted a smaller case and grabbed the small rosewill but it does not fit a hyper 212 cooler. I am using a silenx cooler from newegg. It is a 92mm cooler.

www.newegg.com/Product/Product.aspx?Item=N82E16835226050

It is not as good as the evo cooler.
I got 68c at a 4.2 clock with 1.26 volts.



This setup I am getting 72c to 74c at 1.22 volts with the 4.2 GHz oc.

I was hoping to go lower like 1.19 or 1.20. To keep heat down as much as I can.

 

[bryanW1995]

There is a huge volts difference btw 4 and 4.2 and btw 4.2 and 4.4. It is like nearly 0.075-0.1V for every 200Mhz at this level
And beyond that it only gets worse.

A proper overclocking graph is roughly linear at low overclocks until it hits the wall from my experience. So you likely need around the same amount of vcore increase to go from 3.8 to 4.0 as you do to go from 4.0 to 4.2, but going from, say, 4.4 to 4.6 is probably getting onto the cliff for most chips.

The way to truly measure is to run stability tests at low voltages on stock speeds to find your min vcore at stock, then inch vcore/clocks up in small increments. That's a real PITA and takes a lot of time, so what most of us do instead is see how high we can get our clocks on stock vcore, then measure the increased vcore required to get higher clocks.

 

[borisvodofsky]

Doesn't make sense to under-volt. the power-saving is negligible because you're likely running idle most of the time.

As for increasing the CPU longevity, that's also silly, because 3 years is all it's good for before you'll need an upgrade anyway.

And we're already quite sure that the 2500k can tough out at least 3 years at 1.35-1.4 volts.

 

[philipma1957]

Doesn't make sense to under-volt. the power-saving is negligible because you're likely running idle most of the time.

As for increasing the CPU longevity, that's also silly, because 3 years is all it's good for before you'll need an upgrade anyway.

And we're already quite sure that the 2500k can tough out at least 3 years at 1.35-1.4 volts.

I would normally agree. I did not bother with the larger case and larger 120mm air cooler. I was looking for the low voltage to keep heat down since the case is small and the cooler is a 92mm air cooler.

This under volting is keeping me under 76c /

If I ran at 1.30 volts I was over 83c.
I suppose I could hunt for a better low profile air cooler. I only want 4.2 clock and I am getting that. This machine is quiet small and powerful.

 

[XtremeCuztoms]

my 2600k daily runs 4833.03 24/7 @ 1.41v
temps on water are 22-28c idle and 64-66c load... SB just doesn't handle voltages well for extended periods..

 

[aaksheytalwar]

my 2600k daily runs 4833.03 24/7 @ 1.41v
temps on water are 22-28c idle and 64-66c load... SB just doesn't handle voltages well for extended periods..

You got low volts. High speed. Low temps. What is the problem?

 

[Don Karnage]

Low Votage? Who cares

 

[philipma1957]

Low Votage? Who cares

yeah but you have a beast!!!

 

[BD231]

These chips are voltage sensitive in both directions. Give them to much and your performance will go down, give them to little and it'll go down as well. Best to test via something like Cinebench to ensure you're still getting the correct performance out of your chip at lower or higher voltages.

You can fit a full height tower cooler in your case by the way, I get better temps at 4.7 so it's really the last thing you need to max it out. I can turn my cpu fans off and run fanless at 4.7 :awe:. 480 kinna kills it tho ... if only water cooling wasn't so frikin loud, I'd jump on a water block.

 

[Kenmitch]

I got the numbers down to 1.22 volts and it runs well.

at 1.20 volts it does not boot.


I wanted a smaller case and grabbed the small rosewill but it does not fit a hyper 212 cooler. I am using a silenx cooler from newegg. It is a 92mm cooler.

www.newegg.com/Product/Product.aspx?Item=N82E16835226050

It is not as good as the evo cooler.
I got 68c at a 4.2 clock with 1.26 volts.



This setup I am getting 72c to 74c at 1.22 volts with the 4.2 GHz oc.

I was hoping to go lower like 1.19 or 1.20. To keep heat down as much as I can.

Did you stress test it? Seems strange it won't boot at 1.2v's but 1.22v's is good to go.

 

[philipma1957]

I ran the prime95 at 1.22 volts and it worked well.

then lowered the voltage setting to 1.2 and went to a black screen when I went to boot. I am going to play around for a while.



I am trying to balance noise/heat/size/performance with this build.

wei is 7.6 for the cpu. 7.8 for all other ratings.

It is a nice blend of the above goals.



As for a quiet htpc my i5 2500t build is everything I can hope for. same case

http://www.newegg.com/Product/Produc...82E16811147123

and this gpu

http://www.newegg.com/Product/Produc...750%20ultimate

with this cpu


http://www.provantage.com/intel-cm8062301001910~7ITEP3FE.htm


a pair of crucial m4's some good fans


and you get a close to silent htpc that can do some decent gaming.

wei is 7.3 for cpu 7.4 for gpu and 7.8 for ram and ssd.


this build is super good for power.

30 to 40 watts for surfing the net. 50 watts for netflix and 65 watts for a bluray .