My new 2500K @ 5.0ghz

3t3P · 28 Sep 2011 at 11:00

Hi all I was doing a late night ocing session with my new hardware and got to 5.0ghz but it needs 1.5v core and don't know if it's fully stable yet as have only done five mins each on Prime 95 and OCCT.

I know there's a sticky thread stating 1.35 is a safe upper limit but found this doesn't reallly get me anywhere.

I'm on a watercooled loop and temps max at 71-72 degrees.

Thoughts?

Plec · 28 Sep 2011 at 11:20

3t3P said:
Thoughts?

Personally, i would reduce your clock by ~400MHz and focus on getting it stable with reasonable voltages (under 1.4v) and then build your clock up incrementally from there (if you feel you need more than 4.6GHz).

3t3P · 28 Sep 2011 at 11:54

I think I am at least 4.5ghz stable at 1.425V

I was just wondering if any more data of 1.5v core was available in terms of degradation/fried CPU's.

You see some people rocking these volts...

StonedPenguin · 28 Sep 2011 at 11:58

Shouldn't need 1.425 for 4.5ghz fella, I'm at 4.4ghz and have voltage on auto (offset mode +) and on load it gives a vCore of 1.296 > 1.304 and peaks at 1.312 occasionally, fully 100% stable.

3t3P · 28 Sep 2011 at 12:02

The 1.425 was from an Asus ocing guide on another forum. It was a suggested starting point I guess. Overall my chip probably isn't 'the best'.

Gonna see if 1.49v @5ghz is stable

reflux · 28 Sep 2011 at 12:03

I'm at 4.6ghz with 1.35v, but I have a 2600K

Azarathal · 28 Sep 2011 at 12:17

Mine is 4.4ghz @ 1.34v, 4.5ghz needs about 1.37v :S Make me wonder whether I've got a not-so-good chip, I'm more than happy with 4.4 though.

Leeroy123 · 28 Sep 2011 at 12:18

Every chip is different 5ghz for 1.5v is ok, use offset so it isnt 24/7 though

wildman · 28 Sep 2011 at 13:42

I thought 1.3x was the max vcore before you started frying ? unless you are using WC then i guess the temps would be ok as I assume max safe voltage is based on how much heat is generated and how fast your cooling can dissipate the heat before stuff starts to melt ?

Marine-RX179 · 28 Sep 2011 at 13:47

Unless you are playing on CPU demanding games that don't use all four cores (i.e. WOW, Total War series), then there's not much point clocking your CPU that high and shortening its lifespan, especially when it won't really give you any extra performance in games with you only got a single overclocked GTX470.

In most games, the frame rate for your GTX470 would be the same may it be the 2500K at 4.0GHz or at 5.0GHz...so why risk killing your CPU?

Sin_Chase · 28 Sep 2011 at 13:51

wildman said:
I thought 1.3x was the max vcore before you started frying ? unless you are using WC then i guess the temps would be ok as I assume max safe voltage is based on how much heat is generated and how fast your cooling can dissipate the heat before stuff starts to melt ?

Nope.

Temps are an element of longevity but that's not to say you can pump 3volts through a chip and keep it at a suitable temperature and it not go pop

3t3P · 28 Sep 2011 at 14:05

Interestingly the 2500K at 5ghz draws about 20 watts less in-game than the Q6700 @ 3.9ghz did.
450W and 470W respectively

What is really interesting is at desktop before the Q6700 setup drew 270 watts!
This sandy build is 170 watts!
That's an entire lightbulb right there!

Yes I realise temps aren't everything when pushing more volts through, but I want to use my watercooling gear to its potential.
I also realise that 4.5-5ghz with my GPU there is nothing in it and I should be sensible...

It's just nice number I've reached, it was a target of mine too and wondered if the chips are ok at these stress levels.

It seems game-stable atm at BIOS 1.49v (just completed spacemarine which DID BSOD yesterday when it was unstable),
withCPUz saying a droop down to 1.47 during 100% load

I'll try 1.485v in bios next

edit 1.485v failed prime 95 after four iterations

Sin_Chase · 28 Sep 2011 at 14:43

5GHz at those volts is not a 24/7 clock to be fair.

Aim for the best clock you can get at 1.35 and be happy knowing your WC is making the best of the heat generated.

I get the need to go more, mines at 4.5GHz and I want more but actually have no reason for it

3t3P · 28 Sep 2011 at 14:55

Yeh I'll see how low I can get the volts back on 4.5ghz

It does seem quite stable at 5ghz with 1.49 atm

Gratuitous shot of the new gear in situ

:

night shot:

Sin_Chase · 28 Sep 2011 at 16:05

Looks good dude.

When I have some extra cash at hand next year I am very tempted to do a sick custom loop on CPU/Whatever new GPU I end up with

Robertnpmk · 28 Sep 2011 at 16:16

That looks awesome, going to read your buildlog now.

I have my 2500k at 4.5 Ghz prime stable at 1.34 Volts, As that is the highest threshold most guides tell you to go to Im happy with 4.5 Ghz. Never noticed any lag so wont push it further.

Bleek · 28 Sep 2011 at 17:17

I'm not really in to all this case modding (I have a black case hidden away) but I have to say that does look rather good.

joeyjojo · 28 Sep 2011 at 18:30

I was curious about how much power these chips might be using at these OCs. The following assumes 3.3 GHz at 1.17 volts has a TDP of 95W (which is a worst case scenario remember).

4.5 GHz at 1.34 (as in Robertnpmk) is a 36% clock speed increase for a 79% power increase (169.9W).

5 GHz at 1.49 (as 3t3P) is a 52% clock speed increase for a 146% power increase (233.4W). 3t3P did you measure 170W at the wall under load at this clock?

Diminishing returns definitely

3t3P · 28 Sep 2011 at 19:18

Thanks guys, I love staring at my PC hehe

joeyjojo how do you work out the wattage increases? multiplying the OC amount by the 95W TDP?

The readings I gave are at the wall yeh
170W idle
450W load

So no one has really been happy with my 1.49v lol

I'll scale back to 4.5ghz like the whimp I am tonight....

Leeroy123 · 28 Sep 2011 at 21:13

Its down to personal preference, im happy to use upto 1.5v. Doesnt bother me as long as the temps are right