What is your official opinion on "Prime Stable"?

[Matt2]

Well, in my endeavor to hit 2.7GHZ on my Opteron 170, I have hit a wall at 2.6GHZ. I can run 2.6GHZ @ 1.4v Prime stable, but 2.7 @ 1.54v and everything in between fails prime on Core0, Core1 is fine.

Despite the prime failures after ~2.5 hours, I proceeded to game at 2.7ghz. Every game I have played has run stable, 3dmark06 is stable.

I was just thinking that P95 is a highly synthetic benchmark that produces a prolonged load enviornment that will probably never be replicated in real world usage.

That being said, think it's a good idea to disregard Prime as long as the CPU is game stable?

I know this probably seems ridiculous to most of you, but I just want the most performance out of my chip and am weighing my options.

Thanks

 

[Fallengod]

If it fails in prime, that means "at some point in time" it will fail in windows doing whatever would duplicate the scenario that failed prime. I dont see why youd wanna take a chance. You know how small of a difference 100mhz makes? You wouldnt even notice it in gaming at all....Maybe in encoding, like 1 min faster lol... Keep it prime stable. Of course, its your computer.

 

[biostud]

Originally posted by: modempower
If it fails in prime, that means "at some point in time" it will fail in windows doing whatever would duplicate the scenario that failed prime. I dont see why youd wanna take a chance. You know how small of a difference 100mhz makes? You wouldnt even notice it in gaming at all....Maybe in encoding, like 1 min faster lol... Keep it prime stable. Of course, its your computer.

my words exactly.

 

[Makaveli]

I think if u can make atleast 12 hrs prime stable. she should be good to go. on both cores of course.

 

[nealh]

while I agree to a point ..there are so many variables that affect a windows environment..it is not always a cpu failure...it ca nbe ram, videocard..etc

once any software app is added so are dlls which can lead to issues..so I blelieve each individual has to decide what he/she does with their computer and what tolerances they have for a possible issue

I always give the example of a computer setup at work..2 offices are running Dell computers(different models) that use Digital angiogram software..they are preconfigured to work with specific hardware..not overclocked in anyway...they setups run win2k and are so unstable..they crash often with numerous issues that in theory should never develop given the software maker picks the hardware and configures everything...this is mission critical to my work....

My home PCs that are overclocked are more reliable and stable...

For me

I like 6hrs of memtest, 12 hrs-15hrs(with no errors..I have to stop prime 95) small and large fft

In my case..this has produced the most stable and reliable pc for thelast 2 yrsI have owned

 

[Unkno]

i believe stability is a MUST. I can't stand having the computer freeze when doing anything. IMO, 24hours is a good way to test if the system is stable. If you failed prime95, it means your system has a chance of crashing even when not running intensive programs (larger chance of crashing when running intensive programs)

 

[KoolDrew]

Originally posted by: modempower
If it fails in prime, that means "at some point in time" it will fail in windows doing whatever would duplicate the scenario that failed prime. I dont see why youd wanna take a chance. You know how small of a difference 100mhz makes? You wouldnt even notice it in gaming at all....Maybe in encoding, like 1 min faster lol... Keep it prime stable. Of course, its your computer.

Ditto. Your overclock should be prime stable. However, just because it is prime stable does not mean it is completely stable.

 

[Diasper]

Use S&M if you want the ultimate and quicker test of your CPU - it heats up more and faster than any other program.

Otherwise OCCT is a good fallback options and faster than Prime too.

 

[lopri]

Well, thing is this. Matt2, it's the question you should ask yourself. In fact, most people who use computer don't know what Prime95 is, and they just use their computers everyday. Sure, sometimes system hangs or reboots, but all it takes is usually, well, a reboot to get back to their business. "If your system is not Prime-stable, it will eventually fail sometime even though it looks stable now" well, that might be true but system instability can caused by other things than hardware itself. Buggy software, spyware, or even your local electricity supply (lol) can cause instability of your computer. If you think your OC is stable enough, then you should be happy with it and enjoy your games. Yes, it can crash sometime but it'd be hard to tell if it's because the system isn't prime-stable or video card drivers are immature or anything.

But things change once you (just like most of us) start thinking OC'ing is a goal, instead of means. Indeed that's why probably you(and we)'re spending more time here @AT, instead of spending time time @your favorite game's forum. And for us to communicate, we need a base - mutual agreement. Without it, we can't communicate. If you think playing games smooth is good enough for stability, it's quite valid. As long as you don't try to force that as a *standard* to others. Some might totally disagree, and some might ask "which games"? And they're all valid within their reasoning.

Just like everything else in the society, it takes time and much "validation" process for something to be a standard. I don't know how, but somehow the Prime95 has become the standard for CPU stability, and I don't think it just happened overnight, by someone or a few people. It acquired its current stance through many years, backed by many people. Will it remain as a standard in the future? I don't know. Probably not. But it sure won't be easily gone in the near future. Until then, you can OC your proc all you want and play all your games. If it does what it's supposed to do, what do you worry about? But the moment you come here in public and say "My stable OC", expect to hear "Let's see the dual-prime results".

 

[Absolute0]

about 4 hours of Prime is good for me. I run the system 24/7 and as long as i dont notice any crashing i'm down with it. I think the whole 12 hours versus 24 hours etc. etc. is totally over rated. At the speeds i run i'm so temp dependent that Prime failing is probably more linked to the window being closed all day adn the processor getting too hot.

 

[potato28]

72 hours minimum for my maximum oc. I use OCCT to find out my max oc.

 

[ericeash]

i use dual prime for 12hrs, but that is only because i never know what i'll be doing with my computer. currently i'm running some extremely CPU instense programs for filtering dvd movies trying to get the best picture possible. sometimes i encoded movies, sometimes i game, and sometimes i just like to surf. but if you only game, and you never see your CPU usage go to 100%, ever. then i think prime for 12hrs is moot in your case. i suggest SNM. it is a really easy program to run, and can stress the CPU harder than Prime95, if you want it to. but it also has a setting to see if you are 75% cpu stable. if your games and apps are not going past 75% CPU usage, ever, then try and see how long SNM can run without error at 75%. SNM also tests the rest of your system including PSU and memory, so check it out.

 

[Matt2]

Yeah I tried S&M, I'm a bit confused. DO I need to loop it and run the CPU tests multiple times for accurate results or is one run enough?

I also noticed that S&M runs the CPU about 3-4*C higher than Prime.

 

[dug777]

Don't care about prime providing i got absolutely no crashes gaming/encoding...it's not as if i'm coordinating nuclear strikes from my home, or decoding the human genome

 

[Fallengod]

Dont lie dug!

 

[dug777]

Originally posted by: modempower
Dont lie dug!

 

[n7]

I'm would agree that P95 stability is really not that important.

I ran my X2 3800+ @ 2.6 GHz for a long time, & though it couldn't pass any stability tests, it never crashed while gaming, or even when gaming & encoding video simultaneously.

I've since dropped it down to 2.4 GHz, since i really don't see the difference, & it's SP2004 stable...though really, who teh eff cares.

You should run yours @ 2.6 man
It's obvious that's its limit, since you have to crank up the vcore to get that minimal improvement.

 

[Hacp]

At a certain point, you will not see a difference. I had an X2 that would run at 2.64. I ran it alot of times at 2.4 and didn't notice alot of difference. Encoding times might have been faster by a minute or two, but thats all. Def not worth the risk.

 

[lopri]

Well.. I can *feel* the difference between 2.6GHz and 2.8GHz with my X2. Even on Windows Desktop.. I'm sure there is a psychological factor, but the difference is there. At 2.8Ghz it's like everything's flying. For example when you double-click on a folder, it feels like popping up even before I finish the second-click. I didn't get that feel @2.6GHz. Once I've been used to 2.8Ghz, even with much lower vcore/temp (My X2 does 2.64GHz @Stock vcore) I can't seem to go back to 2.6GHz.

 

[SynthDude2001]

Originally posted by: dug777
Don't care about prime providing i got absolutely no crashes gaming/encoding...it's not as if i'm coordinating nuclear strikes from my home, or decoding the human genome

I agree. No crashes in any of the programs I use = stable for me. I didn't buy my CPU to run Prime95 day in and out...

 

[Griswold]

Originally posted by: lopri
Well.. I can *feel* the difference between 2.6GHz and 2.8GHz with my X2. Even on Windows Desktop.. I'm sure there is a psychological factor, but the difference is there.

Thats a line for the humbug of the year award. How can you possibly tell or feel the difference of 200MHz while looking at the desktop, while your box is most likely doing nothing? Yes, XP is a bloated resource hog, but not to that degree.

 

[n7]

He's got a good imagination

 

[atybimf]

12 hours dual Prime95 blend test.

 

[kyparrish]

12 hours of in place large FFT's on both cores