i2500k max voltage

[Mark617]

What determines the ceiling of voltage I should pump into this chip for an OC - 24/7 safe use for gaming.

Is it strictly a temp thing? Currently, 4.7ghz at 1.45 volts (need to make this more precise, could maybe drop .05-.1 volts) on p95 blend tests - max temp of around 58 C in an ~ 70 F ambient room.

Water cooled (cpu only)

Furthermore, if I can achieve this speed only touching vcore, is there any reason I need to mess with LLC / PLL / over voltage stuff?

EDIT - Furthermore, I can get a much lower vcore setting with load line calibration at 6 (1-10 settings) but I have read differing views on this setting. Is it again a temp thing? Is the excess voltage at idle that bad if temps are low? I would imagine putting it all the way to 10, or even high at 7/8 would net me even better results. Opinions?

Also, PLL overvoltage doesnt have an on off. It has auto, + or -5% and +10%. Mainly worried how much is too much here and really above !

 

[bennoculus]

I wouldn't go above 1.4 for a 24/7 overclock. Temps need to stay in the 70s if you want to be safe, and 80C is your max/if you don't care as much .

It's not exactly a temp thing, but at the same time it is. You don't want to go above 80C. So you do have some headroom. I'd say 1.5v is your absolute max. It'd be very risky doing so, however.

Personally, I would set the PLL to 1.775 and leave the rest alone. I would also leave PLL overvoltage on auto.

 

[turbohans]

1.45v is nuts for only 4.7ghz, my 2600k does 4.7ghz @ 1.38v. Being over 1.4v makes me worry, and I run into the thermal barrier around that voltage anyway.

According to the manual the max is 1.52v though...

"Supporting Intel®
Core™ i7, i5, and i3 Desktop Processor Series
Supporting Intel®
Pentium®
Processor G800 and G600 Series" << I guess all those apply to the diagram below?

 

[EarthDog]

Every chip is different turbo.

I wouldnt put more than 1.4v for 24/7 personally...temps under control or not. Im a bit conservative, but...

 

[turbohans]

Every chip is different turbo.

I wouldnt put more than 1.4v for 24/7 personally...temps under control or not. Im a bit conservative, but...

I agree! I thought that picture there was only for the 2600k's too, but in the intro it says for I7's I5's and I3's (this is where I got it) soo I am starting to doubt if that would be a reliable source,, but, it is Intel?

The clause to that rule however is that a processor not exceed TJ-max though..

 

[EarthDog]

Its absolutely baffling to me as well. I would NEVER pump 1.52v 24/7 into a chip. I have experienced, what I think was degredation, around there under water. I dont know.

 

[turbohans]

Its absolutely baffling to me as well. I would NEVER pump 1.52v 24/7 into a chip. I have experienced, what I think was degredation, around there under water. I dont know.

For sure! I don't get it! Intel saying you can throw 1.5v at a chip is like Chevrolet saying Corvette's are perfectly safe to drive at 300mph.

 

[EarthDog]

Heh, well, 150MPH since the chip (and vette) can actually reach those speeds (voltages)...but yeah. LOL!

 

[freeagent]

You can feed 1.5 to the chip if it wants it/ needs it.. Like already mentioned, every chip is different. Some require more voltage then others, some require alot more lol. I have no qualms putting 1.5 into mine, but it is a pig, and I am somewhat jealous of other examples that have come off the press

 

[Fr1nge]

You can feed 1.5 to the chip if it wants it/ needs it.. Like already mentioned, every chip is different. Some require more voltage then others, some require alot more lol. I have no qualms putting 1.5 into mine, but it is a pig, and I am somewhat jealous of other examples that have come off the press

I agree, I've feed 1.65v into my 955 under water and it didn't do nothing. I didn't even see long term damage even though I ran it like that for over 3-4 months. Every chip is different, if it wants to blow up at 1.65v it'll do it, if not then your good.

 

[EarthDog]

We are talking about intel here, not amd. BIG difference. 1.65v on you 955 is around .25 over stock isn't it...which is nothing vs .45+ over stock of a 2500k.

Be careful on that advice!

 

[Woomack]

Intel says that max 1.52V VID is safe ( not voltage ) ... while you are using all power saving options and then ~1.5V almost never happens. So you will end with ~1.1-1.45V for 24/7. In real Intel didn't test SB above 1.45V good enough to make official statement so we got only some info that they tested some SB and they think that 24/7 voltage ~1.4V shouldn't damage cpu ( depends who said that it's 1.38-1.42V, at least from what I found in the net ).
As some already said in this topic all chips are different, some will run fine on 1.5V and some will start to degrade at about 1.4V.
I had 3x 2600K. One got max ~1.7V for tests and was running fine for next 2 months without degradation till I sold it, one died after 2h random tests on 1.5V and one started to degrade after 3 months on ~1.37V while was running at about 1.52V for short tests once or twice when I got it. Last one started with stable 1.24V 4.5GHz or 1.28V 4.7GHz and after ~3 months it needed 0.07V more for the same settings.
Finally I sold SB and I will save some money for X79 ... now using AMD ... my friend would add in this place

 

[Mark617]

Thanks for all the information and advice guys.

Here is the weird thing, when I say 1.4xx volts its only under load. Usually it drops .04 volts when not under load due to what I believe is the multi-step load line calibration.

The other thing is as follows: I can run the chip at a vcore of 1.275 - load line level 10 which shows a bios voltage of ~1.375 and it is stable. At the same time, a load line of say, level 6, fails quickly under prime at the same bios voltage (higher vcore - lower load line). I have read a few articles which state a LLC of 10 on gigabyte boards can send funky voltages and so it isnt recommended.

Personally, I would rather run a lower vcore with a maxxed out load line. I feel like this way I get the best of both worlds. A higher stable OC with lower overall operating voltage 24/7 (lower idle volts etc).