• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

How safe is it to stay on an unstable overclock ?

bb gun

bb gun

bb gun

Associate
Joined
29 Jul 2015
Posts
20
I'll try to keep this short, but I would appreciate it if you folk could read it to understand my situation.

After years of owning my i5 2500K and running it at stock, I decided to give it a spin and OC it to 4.5GHz since everyone says it's easy. I do that following this guide (http://www.overclockers.com/forums/...Beginners-How-to-set-your-25-6-700K-to-4-5Ghz) , and right now after all the stress tests and everything the CPU is sitting fine at 4.5GHz, 1.35V.

After that I decided to push the volts lower and see how far I can go. I managed to go to 1.34V where all stress tests still passed and where temps are safe, but when I did the same thing at 1.33V and lower only 1 program gave me BSOD - Intel Burn Test (at Maximum).

The following programs are what I use for stress testing and all of them pass at 1.33 or lower voltage except IBT:


- Prime95 v26.6 (Small FFT's and Blend mode, 6h)
- Intel Burn Test (Maximum 10 runs, Very High 10 runs)
- Intel Extreme Tuning Utility (30 min.)
- AIDA64 Extreme (CPU, FPU, Cache) (1-2h)
- Asus RealBench (5 runs)
- Cinebench (5 runs)
- Battlefield 3,4,1 (2h+)


So far I'm at 1.32V and I pass all the tests stated above with the exception on IBT with full RAM usage, and if I actually set custom settings in IBT where my RAM is 90% of available it ALWAYS passes, but never when maximum ram is used. Now the temps get significantly lower when I reduce the voltage's and I'd rather like to keep it that way.



Now I wan't to ask you guys is how safe can it be to stay on an overclock which only fails at 1 stress test program, because the temps from 1.32-1.34V are quite noticeable.

This PC is used only for gaming and not any heavy encoding, folding etc..., so this is the whole point of starting this thread to see if I should bump the VCore up just because I fail on 1 stress test

My PC specs if anyone is curious are:

i5 2500K
CM Hyper 212+
AsRock Z77 Extreme4
2x4GB RAM
GTX 980 Ti


Thanks in forward.
 
The best stress test is always just doing what you normally do. The risk is of course if you're in a tense situation in a game and your overclock gives out.
 
AiiR said:
The best stress test is always just doing what you normally do. The risk is of course if you're in a tense situation in a game and your overclock gives out.
Click to expand...

This is interesting actually. On my system, I can get 3466 ram to pass HCIMemtest 1000% but not 3600. However, I cannot get a single other program (inc Aida64 stress test and Prime95) to crash using 3600.

This also leads me to believe a lot of people are running ram that can't pass HCIMemtest but otherwise their systems are perfectly stable in everyday usage.

For my old 4770k I ran it daily at 4.8ghz, but it would fail Prime95 - that system never crashed otherwise.

gavinh87 said:
Or it corrupts a file and has you installing windows again.
Click to expand...

That's what system image backups are for :)
 
Last edited by a moderator:
Rroff said:
Or corruption creeps into saved games, etc.
Click to expand...

I'd say you'd notice other symptoms of an unstable overclock before you'd notice file corruption. File corrpution is usually the last issue that rears it's head when pushing an overclock.

On the flip side, it's only a save game, not a space shuttle launch :D
 
amigafan2003 said:
I'd say you'd notice other symptoms of an unstable overclock before you'd notice file corruption. File corrpution is usually the last issue that rears it's head when pushing an overclock.

On the flip side, it's only a save game, not a space shuttle launch :D
Click to expand...

If you are at the point it is passing most stress tests but only failing on extended memory testing, etc. then the most likely issues are corruption in stuff like large file downloads, saved games, etc.
 
I would approach the the other way around. The temp difference may be noticeable between 1.32 and 1.34v but you’re not running IBT in real life so it doesn’t matter. If your load temps are reasonable outside of synthetic tests I’d go for stability and up the volts.

Also it’s an old CPU, by the time you kill it from volts/temps it’ll be worth 20 quid so who cares...
 
No game will come close to IBT levels of load. You'll be fine as long as your games are stable.
 
I think the difference between IBT and the others is that IBT is using all your ram whereas the others tend to be quite light on the ram front.

However IBT could very well be complete overkill for a gaming system. Personally I use HWBot's x265 bench tool in 4K 'overkill' mode as I actually do render x265 footage and it uses up all my RAM plus would want more.
 
When I had my 2700k I ran it with 1.368v for 4.8GHz, under custom water. I found anything beyond 1.425v created insane temps.

Not sure what cooler you're using but as suggested a clean and a remount may help. Or an upgrade to an AIO.

Definitely leave the ram alone, in terms of corruption ram tweaking is the worst offender. Any gains to be had just aren't worth the risk.
 
Ok, thanks for all the replies guys.

I think I might revert to 1.34V to save myself of starting new games in games I play, but do you think for an i5 2500K @ 4.5GHz that temps between 60-70C when gaming is high ?

Thanks.
 
You must log in or register to reply here.
Back
Top Bottom