• 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.

Is prime95 realistic?

I mean will the cores really work that extreme? With gaming

:)

With gaming, nope nowhere near really. It is generally used to prove stability, then people like to post screenshots to show how big their e-peen is :D
I'm sure there are better ways of testing other than leaving your CPU screaming for hours on end.
 
The best stability test is using your pc, prime , realbench , Ibt etc none of them will be as good as everyday use, for example i could pass Aida and realbench for 12 hours with no issues yet gaming and simple web browsing would sometimes cause a hard lock, a simple voltage bump sorted the issue.
 
Prime95 is great for testing stability if you want to use your pc for running Prime95.
 
Prime95 does not prove stability, all it does is boil the CPU, you can run Prime95 for hours and then BSOD as soon as you load a game up.

Asus Real Bench is much better.
 
Good thing that these days you can set a known stable overclock level for both then.

AVX offset is a thing and exists.

I know that in Prime95 the 7700k I have is unstable above 4.7, checked with over 13h overnight
I also know that in Realbench it's fine at 5.0, checked with 8h during the day following

AVX offset set to 3 (=300 MHz) and I'm happy in my own mind that I'm not ignoring something the cpu can do and might use at some point in the future even if I don't use it at the moment.

Hardly takes effort, you need to sleep sometime, most people also go to work, run the bench test then.

Not to mention, a stock cpu can pass Prime95. Doesn't feel right to overclock it so it can't.
 
what a load of BS

Why is it a load of bs ? testing with prime only proves stability in prime , you can test with prime for 24 hours then as soon as you decide to load up a game it crashes or five minutes in and it crashes. Maybe humbug should have said prime does not prove real world stability.
 
Prime is the daddy IMO. what will take realbench hours to crash, prime can have you done in 10 mins! BANG!

No messing. If you can run prime 95 for 24hrs, you should have a rock solid platform.

If you cant prime for 24hr runs then your PC is not stable, it's quite simple.

You might think that because it runs cinebench a couple of runs, or 4hrs of realbench your pc is stable. Well it probably is not!! You will find that there is very slight instablity that can rear its head up like once a month or even a year.

Just my 2c.
 
The reason most people use Prime95 to stress test machines is because it's a math program and math is one of the most basic things with CPUs, if it can't math it can't CPU :P

Having said that just because an overclock is stable running one application doesn't necessarily mean it will run a less demanding one (though odds are it will) hence the existence of applications designed to stress a CPU like Realbench, IBT, etc.
 
I use overnight runs of Real Bench and x264 stress test v2 (not at the same time), as well as gaming to gauge stability. The result is a rock solid OC for gaming, video editing, encoding etc., all the things regular people use PC's for.

I'm not a fan of the highly synthetic stress tests. Not only do they stress in a way you'll never manage in day-to-day use, they heat the CPU an unreasonable amount which can destabilize your OC anyway. I think stability is rather subjective and should represent how you use your PC. If your PC crashes while calculating prime numbers, then it's obviously not stable when doing that. However, how often do you leave your PC calculating primes?
 
Last edited:
I use overnight runs of Real Bench and x264 stress test v2 (not at the same time), as well as gaming to gauge stability. The result is a rock solid OC for gaming, video editing, encoding etc., all the things regular people use PC's for.

I'm not a fan of the highly synthetic stress tests. Not only do they stress in a way you'll never manage in day-to-day use, they heat the CPU an unreasonable amount which can destabilize your OC anyway. I think stability is rather subjective and should represent how you use your PC. If your PC crashes while calculating prime numbers, then it's obviously not stable when doing that. However, how often do you leave your PC calculating primes?

Trouble is, if it is not stable calculating prime numbers, then it is not stable.

So you could be gaming or whatever, then all of a sudden, random crash. You might only see that once a month or something, but it is still not what I would call stable for 24/7 use.
 
I'm sure prime can be useful, but I don't think any one application is useful for stability testing.

When I get back to overclocking, my primary will be RB, then overnight encoding, then some gaming, then if I can be bothered, I may run Prime overnight.
 
Trouble is, if it is not stable calculating prime numbers, then it is not stable.

So you could be gaming or whatever, then all of a sudden, random crash. You might only see that once a month or something, but it is still not what I would call stable for 24/7 use.

That's not the case though. Unless you're replicating the same way it crashes in stress testing, you won't see the crash. I've had my 6700k clocked to 4.8GHz for 9 months now, tested with the above methods, never had an issue. As far as I'm concerned, that's 100% stability.
 
You might think that because it runs cinebench a couple of runs, or 4hrs of realbench your pc is stable. Well it probably is not!! You will find that there is very slight instablity that can rear its head up like once a month or even a year.

Just my 2c.

So in your opinion that crash that occurs once a year can only be due to OC and nothing to do with windows or a program ? Once a month the same crash yeah could definitely be a stability problem, needs to be looked at, and can't be concidered stable, but one crash a year could be anything.

That's not the case though. Unless you're replicating the same way it crashes in stress testing, you won't see the crash. I've had my 6700k clocked to 4.8GHz for 9 months now, tested with the above methods, never had an issue. As far as I'm concerned, that's 100% stability.

Yeah 9 months without problems that's stable. Personnally my 4790k craps out on a prime95 overnight test, but has never crapped out in a year, so I wouldn't call prime95 the end all be all of stability for an everyday user, in my type of usage anyways (gaming, encoding, DAW VSTi/recording)
 
So in your opinion that crash that occurs once a year can only be due to OC and nothing to do with windows or a program ? Once a month the same crash yeah could definitely be a stability problem, needs to be looked at, and can't be concidered stable, but one crash a year could be anything.

Yes, that is my opinion. Software bugs are often well known about and can be accounted for most of the time. It's when you get those crashes in software that is normally stable.
 
Why is it a load of bs ? testing with prime only proves stability in prime , you can test with prime for 24 hours then as soon as you decide to load up a game it crashes or five minutes in and it crashes. Maybe humbug should have said prime does not prove real world stability.

no it doesnt. if it passes 24hrs of small ftt AND 24hrs of 4096k large ftt using max ram, it should not crash in a game or anything else unless there was a user error, where the doughnut set a minus offset for idle/partial load :rolleyes:
 
prime is useless, its not a stress tester, its designed to calculate pi, which just so happens to quite demanding but not stressful:
2czoftt.jpg

above over 73hours prime stable.

below, fails in just under 1 hours 15 mins of linpak.
kb7jg9.jpg


I don't bother with prime anymore just run linpak for a week nonstop and then its good to go.
 
Back
Top Bottom