What is "stable"?

spitzoid

spitzoid

spitzoid

Permabanned
Joined
21 Apr 2008
Posts
872
Location
Gloucester, UK
Guys,

Newbie to o/c'ing.

Just wondering...what "stable" actually is or means? Does it mean that your temps are stable (albeit high)? Does it mean that your system doesn't crash under sustained 100% load? Or something else?

Also, what sort of temps would you be happy to accept whilst running Prime95? I put together my first rig last night - Core i7 920, Asus PT6, 6GB Dominator 1600Mhz, Noctua Cooler, Antec 1200, 4870X2, WD Raptor - and immediately o/c'ed it to 3.5Ghz. When I ran Prime 95 it fairly quickly got near to the 70c mark which I wasn't happy with - considering it was running at 29-35c idle - but was wondering if you would be happy/accept those temps and not be a pussy like me and stop it before it hits 70c.

Any advice on max temps, sustained temps, "stable" definitions, good benchmarking software, i7 920 o/c settings etc would be appreciated,

Cheers
Worried noob oc'er
 
stable means it's running without any errors, usually if its benchmarked over night it'll stay stable. although if you're right on the edge of stability even things lke the change of seasons causes temperature changes and could cause it to become unstable.

temperature depends on cpu, cooling and overclock, i personally wouldnt like to run mine higher than 50'c

Most cpus these days shut down when they get too hot, check the bios there should be options to select warning temperature and shutdown temperature.

MW
 
Pilky01 said:
8 hours of prime 95 normally means its stable
Click to expand...

Not true, I had my cpu overclocked and was fine running prime for over 48hours (was away for the weekend so thought good time to test!). No errors or anything.

When trying to play any game my PC would instantly crash and burn. I have now got it to play everything I throw at it, be it on a slightly lower clock, however I class my PC as being rock solid.

Stable = OKish temps, never crashes using day to day activities that you use your PC for.

Edit: Temps for me dont go over 50 on full load (e6300), soon to be going under water and new ram to push it further, however I would be happy for it to run full load at 60. On a day to day basis it is very unlikely my PC would be stressed 100% anyway.
 
Last edited:
Goog god, stable means your overclocked computer runs exactly the same way and does everything apart from being a little faster than if it was not overclocked.
 
stable means it dosnt crash at all under your personal work load....most OCers use the 8hr prime test but tbh 2 6 8 12 hrs what ever you have time for and are happy with is stable.
also try intelburntest and OCCT.
i had a system that wouldn't last 5 mins in Prime but never crashed under normal usage...so for me it was "stable"
 
As long as it doesn't crash with the things I want it to do I class it as stable, only ever ran prime for just over an hour before, never had any problems.
 
Sayso said:
stable means it dosnt crash at all under your personal work load....most OCers use the 8hr prime test but tbh 2 6 8 12 hrs what ever you have time for and are happy with is stable.
also try intelburntest and OCCT.
i had a system that wouldn't last 5 mins in Prime but never crashed under normal usage...so for me it was "stable"
Click to expand...

That's perhaps your definition of stable, but it's not the accepted one.

Stable means it doesn't go down under load.

It's not exclusive to Prime 95 (being pedantic you should use Orthos), Furmark, OCCP, looping 3dMark, gaming whatever. If it fails any of those then it's not stable.
 
Mines 8 hours 54 minutes Prime 95 Blend test stable.

But the true test will be if it can run my two FAH Linux SMP clients within two VM's and the GPU client on the 9800GTX.

Uploaded 2 WU's, so far so good :D
 
oceaness said:
Mines 8 hours 54 minutes Prime 95 Blend test stable.

But the true test will be if it can run my two FAH Linux SMP clients within two VM's.

Uploaded 2 WU's, so far so good :D
Click to expand...

Does that mean it failed at 8 hours 54 minutes, or does that mean you set it and went to bed and stopped it 8 hours 54 minutes later? :)
 
Before you test for CPU stability, ask yourself what you normally use your PC for. If you intend to do folding, encoding or anything that uses 100% of your processor on a continuous basis, a long run of Prime is a good test of stability. If you use it for gaming, graphic design (or anything else really) then a shorter Prime run or a stressful bench mark such as those mentioned previously will do.
 
SillyHeatsink said:
Before you test for CPU stability, ask yourself what you normally use your PC for. If you intend to do folding, encoding or anything that uses 100% of your processor on a continuous basis, a long run of Prime is a good test of stability. If you use it for gaming, graphic design (or anything else really) then a shorter Prime run or a stressful bench mark such as those mentioned previously will do.
Click to expand...

********.
 
SillyHeatsink said:
Heh, care to back that up?

I can see where you're coming from, but there's no point in stressing a system for 8 hours at 100% CPU and RAM usage when the vast majority of users and programs never even touch 75% usage on a quad core.
Click to expand...

Ok. The errors don't happen because the system is under stress, the errors just happen. The system being under stress does two things, it makes them more likely to happen, and it also does a lot of calculations and checks every result.

You may not notice when the vast majority of errors occur, and you may only get an error once in a while.

But when your hard disk gets corrupted, or your PC dies in the middle of a bios update or a driver update or in the middle of a document, or while you're downloading a big file, or just managed to finish off the boss with 5% health left in a game that's when you'll regret it.

It also means that any time your PC bluescreens or freezes you won't be able to pinpoint the cause and will have to put it down to the overclocking.
 
The Halk said:
The errors don't happen because the system is under stress, the errors just happen.
Click to expand...

Well, I like to think that I'm man enough to admit when I'm wrong, and I was wrong. Thank you for the explanation, you've taught me something! I was under the misapprehension that increasing processor utilisation caused increasing heat which caused errors, and these errors only occurred above certain temps. Cheers, and I apologise for ever doubting you!
 
You must log in or register to reply here.
Back
Top Bottom