You might remember a while ago I looked into overclocking it more than the 4.2GHz I always had it on. I played with it, but wasn't seeing a difference in FPS so left it at 4.2GHz. But since I'm looking to upgrade now, I may as well overclock as far as I can in the meanwhile. I was wondering if there are risks, outside of the CPU degrading faster. I wouldn't want to short out something else or screw some sensor up or not be able to boot again. The highest I've tried so far is 1.45v (temps were below 80C on air when playing). Intel says the limit is 1.52v, but don't know what that means since higher voltage = increased rate of degradation rather than instant death afaik. How far would you go?
Highest I took on my 2700k was 1.7 on water but that was for benching. Sat at 5.2ghz and 1.5v for two years with no sign of degradation.
What are you cooling it with? I'd you're cooling it sufficient you should be able to get 4.8-5 without too much trouble, unless you have a full on dog of a chip or board. With overclocking there is always a chance something could go pop. If you're going to REALLY push it, just be mindful of what else you have it the system for if it goes bang.
Yeah very much so. With Sandy it wasn't the chip that will fail. It was the board VRMs. Those work the hardest when you up the volts. That said, with it being as old as it is I would probably push it to breaking point
When I first built the PC, I just went into the bios and used the auto setting for 4.2GHz. I tried the 4.5GHz one, but it BSOD. I never felt like I needed more, so kept it at 4.2GHz (1.3v). Over time I learnt about CPU usage, bottlenecking etc. and recently looked into it again. Although TBH I'm not sure what a faster speed changes really. A Hyper 212 EVO and the board is a Maximus IV Gene-Z. Currently have it on 4.7GHz at 1.456v after it froze at 4.8GHz but seemed fine at 4.6GHz. How do you define 'REALLY pushing it'? You mean the motherboard could eat it? I wouldn't want that, that's the last part I wanna change really. I wouldn't want to hasten its death anyway.
With a 2500k I doubt it. Maybe more with a 2600k or higher. Overclocking requires one skill, and that's learning not to take the pee lol. Like look up max voltage (I wouldn't give it 1.5 myself) and lower it a little. If you push too hard you'll soon find out, then back it off a little. As long as you don't accidentally punch in a bad number for volts and keep it sensible (I would top out at 1.425v) you shouldn't cause any damage. What I meant when I said VRM was like, long term learned. Obviously any virgin fresh board has new components, but as time goes on it's the VRMs under the most strain if you are heavily overclocked. And it will usually be the board that would fail first.
Yeah - look up some old OC guides for the cpu and your board if available - don't leave any voltage settings at auto.
The board should be good, but you're going to be pushing the limits of the cooler. By really pushing it I meant 1.5v+, 5ghz+. If I were you I wouldn't risk it. That motherboard would be worth a pretty penny on the open market. CPUs are more robust than the CPUs, so you're left with loads of CPUs and not many boards.
Haha we are talking 2011 tech so the CPU and board have been around a while, though I got them over 5 years ago from this very forum. I turned the voltage down now so I'm trying 4.7GHz @ 1.425v. It's not on auto, I've just been changing the multiplier and voltage. Haven't been stress testing as not much point at this stage since I'm hoping to find a non-horrifyingly-priced 3770K/2700K/2600K. There was an exact guide for this board and CPU for 4.4GHz, but I got a BSOD. I suppose if the motherboard did die, that would be my cue to get a new motherboard and CPU completely anyway. But I won't go past or near 1.5v if it might take something else out. I'm guessing there's not that much difference visibly in a few MHz anyway.
I can't remember the voltage but I set up a 2500K on that exact board using the Auto 4.8GHz setting and adjusting slightly (was intending to tweak the settings at a later stage but we never got around to it) and a 120mm Antec Kuhler AIO (can't remember exact model). If you were fussed about keeping the system it might be worth checking out cheap cooling upgrades as an option.
So far 4.8GHz @ 1.45v and 4.7GHz @ 1.41v have failed, but 4.6GHz @ 1.45v and 1.4v haven't had any issues yet. At the time at least, the Hyper 212 Evo was the go-to. I'm not looking to get into liquid cooling though.
Honestly, all in one coolers couldn't be simpler. At the bare minimum, it's just a case of removing the rear exhaust fan and the existing tower cooler and replacing the exhaust with the 120mm fan and rad from the AIO, where it performs a similar purpose whilst at the same time cooling the CPU. You can get them second-hand for buttons these days. I bought a 240mm brand new from Scan for under ~£35.
I looked it up, it's as simple as putting on any other cooler, I'm used to seeing all these pipes everywhere when people talk about liquid. I'll keep it in mind if temps ever feel like a problem.
I wonder if it may be that temps ARE the problem. I'm no expert and frequently solicit advice on here regarding what I think must be magical heresy, but turns out to be stupidity on my part.
The temps never went past 70C I think (80C at 1.45v) so I don't seem to be having any problems there.
Many of the early ones tapped out around 4.7. Unless you had water. I have a review here in CPC of a Scan Scorpion? something like that. It had a 2600k and GTX 590 in it. Even with lots of rads it was still going into the 90s on 1.5v. It was clocked to 5ghz, IIRC. You need pretty extreme cooling to keep one at 5ghz IMO. And you need to give it crazy volts too. I think most backed off to around 4.7.
