I5-2500k OCing seems stupid easy, I gotta be missing something...

doublepedaldylan

Distinguished
Jun 2, 2011
81
0
18,640
So, I've worked mine up to 4.5 GHz and mostly stable (once every few days I get a random reboot.

I'm new to the whole OC scene, and been reading various guides. The thing that I question is, besides a few CPU settings and protections either being enabled or disabled, it seems like overclocking goes as follows:

1) Up multiplier until you BSOD under load.
2) Up Vcore until you can go ~15 minutes on Prime95 blend w/o an error.
3) Repeat 1 and 2 until desired level reached.
4) Go for 6-12 hour runs of Prime95 blend, if errors, up Vcore again until no errors
5) Congrats

Am I missing something here? Is there a point you want to start changing other voltages (besides OCing RAM).
I know it is suggested to lower RAM clock when trying to push for higher numbers.

I'm working on 4.6 now, but Vcore is
at 1.380V (not sure if its bad taste to post settings)
so running out of safe overhead room.
 

HugoStiglitz

Distinguished
nope thats pretty well it.

the K and Black Editions of Intel and AMD CPU's have made overclocking a breeze.

your Vcore isnt to bad.

run the bench marks, drop ur cpu overclock, up ur memory speeds and benchmark it.
work out if ur better off with CPU clock speed or memory speeds.
 
If you want you can try isolation. Set all components to a lower then default setting except the component your overclocking. Then basically raise that device until bsod or failed prime test, increase voltage, ect. Once this is done you will have the max overclock for that component. You can do this for each device and then combine the results one at a time. If something goes wrong, then the device you combined last is the limiting factor of your machine.
 

Frizzo

Distinguished
Sep 17, 2010
599
0
19,010
it hasn't always been like this. they know there is a large tech community, so they work to accommodate us. generally overclocking not only was harder, but the margin of speed increase wasn't nearly as high. the bottom line does stay the same though, if you don't know what you're doing, you might destroy your computer. or at the very least, dramatically reduce it's lifespan.
 

Fortunex

Distinguished
Mar 16, 2009
241
0
18,690


In general that's all you need to do. You can tweak other settings to stabilize higher OCs and work your Vcore down a bit, but those aren't really needed. If you start overclocking RAM or the BCLK you may have to tweak other settings, but generally it's just multiplier and Vcore.

Instead of running standard blend for 15 minutes, run a custom blend using 90% of your RAM, and run FFTs 1344 and 1792 for 15-20 minutes each (set the min and max FFT size to 1344, run for 15-20 minutes, then set the min and max to 1792, run for 15-20 minutes). These will give you a better idea of your OC's stability rather than the first couple FFTs on standard blend.

You can lower your PLL voltage and it may shave off a bit of required Vcore (lowering mine from 1.8 to 1.55 let me take off 0.01v on my stable OC).

You may find some other tips here:
http://www.overclock.net/intel-cpus/968053-official-sandy-stable-club-benchmarks-stable.html
 

doublepedaldylan

Distinguished
Jun 2, 2011
81
0
18,640


I guess from what I read, 1.4 Volts was the 'safe' line, and since I'm new to this didn't want to push it.
Was able to get to 4.3 GHz on stock voltages without any immediate issues, so I was hoping to get a little higher than 4.6 safely.
I'll have to work on memory when I have some time again, school is steamrolling me right now.
 

doublepedaldylan

Distinguished
Jun 2, 2011
81
0
18,640

So I got 4.6 stable at 1.38V (maybe it was 1.385V). Chances I will make it to 4.7 stable?

Also, what does the 1.52V mean. Is this like the "you pass this and your CPU fries" or "you pass this, and you start to gain probability of your CPU frying?" Would you be able to go past with adequate cooling, or are you rolling the dice just due to high voltage f'ing up some of the hardware?

Thanks in advance.
 

Fortunex

Distinguished
Mar 16, 2009
241
0
18,690
1.52v is "go past this and your chip will start degrading very quickly". But 1.51 isn't much different than 1.52, and 1.50 isn't too far from 1.51, and etc., basically, the higher voltage your chip is the faster it will degrade, but the lower the voltage is the less noticable the degredation might be. If you run your chip at 1.55v, it might only last a couple weeks to a month, whereas if you run it at 1.3v it might last 8 years instead of 8 years and 8 months or something.

At least that's my understanding of it, I could be wrong.