• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Prime95, does it matter?

S

sdk

[H]ard|Gawd
Joined
Dec 2, 2003
Messages
1,597
Well I have a question. Specs are in sig. Rite now at 18 x 160 = 2880 it is idling at 37C. I figgure if I went 18 x 167 = 3006 it would be about 45ish. If I ran prime95, I am pretty shure it would fail. But I my old systems what prime95 would fail, would run stable. Just wondering, if I OC" to much, will it slow down my performance?

TY
 
but if it dun't crash on me when I am gaming, does it matter?
 
Originally posted by sdk
but if it dun't crash on me when I am gaming, does it matter?
Click to expand...

The honest answer is no, it doesn't really matter. If you can run through some intense gaming sessions, benchmarks, etc, for a few hours, let's say.....it should be fun. Keep an eye on temps, but you should be okay.
 
If you value performance over 100% stability, then if prime95 fails and your games still run fine, don't sweat.
 
I mean, I dun think I should listen to a program that tells me that I can't get past 18 x 160 when I am running 18 x 167, perfectly stable.

However I do love primmy.
 
Just don't fold or do seti with it not 100% stable... you may send in false results... which is no good hehe
 
If everything runs perfectly except prime95 I wouldn't be too worried, but it does indicate that you've overclocked too much.

I personally will not run a computer for any length of time that is not 100% prime stable, because I do not want to lose work or data due to a crash or data corruption for a few magahertz.
 
To me, stability is the most important factor in a rig. Yes overclocking to get the most for your money is what most people want, but what good would it do if your CPU is returning flase values and you dont even know it. I had a heavily overclocked rig that was gaming stable. When i burn a DVD, the video would would freeze then the dvd program would crash. I burned the dvd 2 more times only to have the exact same thing happen each time around the same part of the video. At the time the rig was not ptime stable. I then turned down the cpu speed by 100mhz, then reburned the dvd. That one came out perfect. I then tested it and it was prime stable. Not being stable for a measly 100mhz isn't worth it to me.
 
I have it at 156 x 18 = 2808. The main reason that 3ghz looks better when running tests, but in the long run, if it burns up my chip. The temps seem to be fine, but my 2700 can't cut it, well atleast I dun think

Could I run 17 x 18 - 3006 and be safe, just so it looks better, or should I play it safe and play it lower.
 
If prime95 fails, your processor isn't doing it's job correctly. I'm going to disagree with the general opinion here and say you shouldn't ignore errors in prime95. Who knows where else those errors will pop up later? Be on the safe side, your not going to get very much more performance AND you get peace of mind.
 
I'm going to say it matters. Even if it's stable in everything else, I've heared of data corruption problems if it's not 100% prime stable (take the sending in false results to SETI for example). It's not worth a few extra mhZ if you ask me.
 
well
a 400mhz overclock should be okay

i guess


i forogt 2 set my vc settings to performance, i got 16.5 and 5900 with it on, doing stuff with it on performance now :)
 
Learn the rule of thumb:

If Prime95 is not happy, MAKE IT HAPPY.
 
I just put together a pc that ran a few open apps, a couple dawn nvidia demos, and 2 instances of cpu burn-in for hours upon hours without issues, yet less than 1 second into prime, it craps out. Im not so sure its without bugs. (yes thats more or less redundant, but i get personal pleasure in making windows cry)

If it doesnt crash, it IS 100% stable, thats what 100% is afterall, a pc that doesnt crash.
 
100% stable means that your computer is working 100% right. If your processor is not processing information correctly (EX: Prime95 FAILS) It IS NOT WORKING RIGHT!
 
prime 95 doesnt run correctly on my stock p3 1.4S - es, but my computer has been running fine for more than a year
 
If you are not running Folding@home or Seti, then failing prime is probably ok. But, if you ARE running those, then there is one VERY simple fact you should consider.


If you fail prime, your computer cannot do simple binary mathmatics without generating an error/getting an incorrect result. This means, that at the fundamental level, your computer is NOT working correctly. You may be stable, but your system is NOT functioning correctly, for some reason. Software, hardware, doesn't matter; your system is not working right. This is binary mathmatics, there is no "kinda right". It's either right, or wrong.

It MAY work acceptably for what you do, may even be 100% stable, but it isn't RIGHT. Take that as you will, just don't run F@H or Seti, since you will return totally invalid results.
 
If Prime95 fails the system isn't stable. That isn't acceptable in my book... but that's just me.
 
Originally posted by SVTLightning
If prime95 fails, your processor isn't doing it's job correctly. I'm going to disagree with the general opinion here and say you shouldn't ignore errors in prime95. Who knows where else those errors will pop up later? Be on the safe side, your not going to get very much more performance AND you get peace of mind.
Click to expand...

yeh exactly. you may think that otherwise its stable, but itll crash and crash HARD sometime in the future...
 
Originally posted by wtf
prime 95 doesnt run correctly on my stock p3 1.4S - es, but my computer has been running fine for more than a year
Click to expand...

then somethings wrong with your system, simple as that.
 
If it's not prime stable, there is something amiss, somewhere. Whether you will ever notice the error/crash is debatable.
 
If I were to run 167 x 18 = 3006
I know that it would not prime95 stable, however if my games were to run good and no random lock ups, would i be affraid of it just going poof one day, assuming that my temps did not cause it?
 
Originally posted by sdk
If I were to run 167 x 18 = 3006
I know that it would not prime95 stable, however if my games were to run good and no random lock ups, would i be affraid of it just going poof one day, assuming that my temps did not cause it?
Click to expand...

Who knows. It isn't working right, though.
 
rite how running spybot s and d, my cpu usage is 100%, when i had 3006, it was like 10%
 
Originally posted by sdk
rite how running spybot s and d, my cpu usage is 100%, when i had 3006, it was like 10%
Click to expand...

What?

You saying that overclocking reduced CPU usage there by 90%? I call BS.
 
Like a car, sure it might run fine for a bit.

But when the wheels fall off in the middle of the highway, you've only got yourself to blame.
 
I wouldnt rely on prime95 telling you your system is stable. Many people have run this app at stock settings with prime95 coming to a halt. I have 2 system now which at stock settings will not finish Prime95. No matter what i change in bios it does not matter. My one system has a mild overclock with 1:1 cpu/memory ratio @ 218 and has never crashed or locked up in the past months it's been running SETI 24/7 while doing video editing/ compression and 3drendering.
Oh by the way for those whom will try to backup their false God Prime95 my specs are as follows:
(2) ASUS P4c800-E deluxe
(1) P4 3.0c
(1) P4 2.4c
(2) 256M Corsair PC3200 XMS
(2) 512M Corsair PC3200 XMS
 
Originally posted by Burning Phoenix
I wouldnt rely on prime95 telling you your system is stable. Many people have run this app at stock settings with prime95 coming to a halt. I have 2 system now which at stock settings will not finish Prime95. No matter what i change in bios it does not matter. My one system has a mild overclock with 1:1 cpu/memory ratio @ 218 and has never crashed or locked up in the past months it's been running SETI 24/7 while doing video editing/ compression and 3drendering.
Oh by the way for those whom will try to backup their false God Prime95 my specs are as follows:
(2) ASUS P4c800-E deluxe
(1) P4 3.0c
(1) P4 2.4c
(2) 256M Corsair PC3200 XMS
(2) 512M Corsair PC3200 XMS
Click to expand...

Stable? Maybe, maybe not. But, it can tell us if one part of your system is functioning within the x86 spec. To elaborate:

If prime fails then your system cannot do basic binary mathmatics. This is NOT something that is up for debate. It is a simple mathmatical function, and your system fails, thus there is either software or hardware that is preventing it from doing math right. Therefore, your system is NOT functioning properly, and you are probably returning erronious data to SETI.

Fix your computer. Just because you don't crash, doesn't mean that it is working completely right. In fact, we KNOW you aren't working right, since there is some basic math (which is what the core operations of the computer are) that it CANNOT do (it fails).

Bad math does not mean your system crashes, or fails. May just be a pixel off somewhere once in a while, BUT... You are running another mathmatical program (SETI). Your system can't do simple math, what makes you think it can do a Fourier transformation?

You cannot complete prime -> your system is not functioning right. Mathmatics are not a false god. They just are. It either can do it, or it can't. Right, or wrong. There is NO gray area. And if it can't, it's broken. Start looking for the cause of the problem, and not blaming a simple test. That's showing you something is wrong. When the doctor tells you you're sick, do you blame him, or take the meds?

Or hell, program the algorithm by hand. It isn't that complex. It's just some arithmetic and some summations. Some arithmetic that your computer can't do. Some arithmatic that a 386 CAN do, and yours can't. Makes you feel all powerful now, doesn't it?

Owned by a 386.

Stop running Seti if you can't run Prime. You can't do math, so DON'T.
 
I will now finish my story since i got one person to reply. First i purposily neglected to add some other details in the previous post.
You people have to stop preaching the Prime95/stable thing because it is not 100% true. I realize some you believe in the idea whom probably will believe in the programmer of the Prime95 is absolutely the most genious programmer out there to date because he made an app on his first try without bugs. Yeh OK maybe you should read all his updates. http://www.mersenne.org/whatsnew.txt
Some seem it may have caused problems running the app to it's entirety.
Anyways what i failed to tell you was last year when i built the 2.4c system which i obviously ran an earlier version of Prime95. It of course didn't finish. I then seen many others complaining of the same thing. I built tens of systems for people last year so i know what i'm doing. Now this conversation has come up in other forums i visit and of course we have Prime95 defenders elsewhere. One person lately mentioned he tried the newest version and now can run prime95 forever. I then downloaded it myself while in doubt it had changed. Well guess what it now runs on that system and even runs while overclocked.
So don't tell me that this app is perfect because its far from it! What bothers me is you guys probably preached prime95 when the previuos versions had bugs and obviously it wasn't running right.
 
Some women say it doesn't matter and talk about it's not the size........it's the motion of the ocean.

Wrong topic......he wasn't asking about "does size matter"....right?

Prime95, it's all subjective. I would say if you cannot run a "burn in test" for a few hours without a reboot, or hardlock then I would say your system is flaky. So yeah, I would conclude that Prime95, toast, sissoft, and any other "burn in test" does matter.
 
all i say is if you get any kind of errors be it prime not finishing, sisoft rebooting, or any other burn in test causing hang ups and what not, do not use the SETI, F@H, PRIME95, etc. type programs. you will return eroneous data and thus, be waisting not only their time but your bandwidth...not to mention putting un due strain on an over worked, unstable processor.
 
Originally posted by Burning Phoenix
I will now finish my story since i got one person to reply. First i purposily neglected to add some other details in the previous post.
You people have to stop preaching the Prime95/stable thing because it is not 100% true. I realize some you believe in the idea whom probably will believe in the programmer of the Prime95 is absolutely the most genious programmer out there to date because he made an app on his first try without bugs. Yeh OK maybe you should read all his updates. http://www.mersenne.org/whatsnew.txt
Some seem it may have caused problems running the app to it's entirety.
Anyways what i failed to tell you was last year when i built the 2.4c system which i obviously ran an earlier version of Prime95. It of course didn't finish. I then seen many others complaining of the same thing. I built tens of systems for people last year so i know what i'm doing. Now this conversation has come up in other forums i visit and of course we have Prime95 defenders elsewhere. One person lately mentioned he tried the newest version and now can run prime95 forever. I then downloaded it myself while in doubt it had changed. Well guess what it now runs on that system and even runs while overclocked.
So don't tell me that this app is perfect because its far from it! What bothers me is you guys probably preached prime95 when the previuos versions had bugs and obviously it wasn't running right.
Click to expand...

If it had bugs, you check to see if others have bugs. If they do, you know it's a software issue with Prime. If they don't, then YOU have an issue. The current version of Prime is about as mature as ANY platform will get. Do you have ANY idea how many people use prime? It's on the order of hundreds of thousands. With that many people, you're ALWAYS gonna find some race condition or mutex that you missed. It's a given fact in Software Engineering. Other than "Hello, World" there is NO program that works perfectly. But you can guarantee it for a MASSIVE majority of the people out there.

Again, you fail prime, there is a 99.9% chance that YOU are the one at fault, not the system.

Stop trolling (the whole "I'll finish and tell the rest now that someone replied" just to get a chance at proving a point is bad taste, if not plain arrogant). If you can't pass prime, you can't do math, your system is broken.

Again, owned by a 386.

EDIT: And did you bother to READ the changelog? 98% of those are performance optomizations, something you ALWAYS do on an algorithm, and 99% of the errors were with multiple tests and memory allocation, or just mem. allocation in the first place. Memory failure is a TOTALLY different error than the prime failing error. Part of me want's to call BS on your not passing/now passing.
 
Stop being ignorant. I'll prove it to you. I still have the 2 versions saved on cd with various software. I just ran both versions on the system i'm typing on. Heres a screenshot. I can run the newest version as long as you want but not the version last year. Explain that one! Oh i forgot the app is perfect and it doesn't make mistakes. It calculates math flawlessly. I'll provide a screenshot of both running the torture too if it helps convert you to being a nonbeliever.
prime95.jpg
 
Listen. I never said it wasn't possible, but you're almost 1:100000 if it was a software glitch, given the amount of people that use it.

I'm a software engineer and a mathmatician. Algorithms sometimes mess up. BUT, with the amount of people testing, and the amount of time their algorithm has had, it is almost foolproof now.

Now then, do you want to prove that there is some kind of error in it, or accept that your system is working fine and take your trolling elsewhere?

I design algorithms. I trust Prime. I know how it works. 99.9% of the time, if you fail prime, your system is borked somewhere, especially from a high overclock.

So, simply, we can assume from that that pretty much, either it runs, it fails, or it crashes with another, unrelated error.

No tool, no program, no test is perfect, let alone in computing. When you get hundreds of thousands of people using ANYTHING, you'll have an error, somewhere, somehow. But, in something this widespread and relatively simple, it's about as foolproof as you can get.

I will never be an nonbeliever. I know how it works, so I trust it. Only someone who WANTS to be ignorant will ignore the simple fact that an algorithm, well tested and designed, is pretty much foolproof, unless you fobar the math.

EDIT: And its still quite obvious that you've never read the changelog or realized what most of the changes were, or have looked at the source for programs like that. Given the number of error correction in something like that, it's almost impossible to break it with out your system being broken.
 
Ok i'm sorry i said ignorant. You sound like a reasonable person . I just wanted you to acknowledge that the app was is not perfect. I would rather hear people say the test is maybe 95% trustworthy and not 100%. It just buggs me that my systems didnt' run the older version without halting and then people demanding i have problems with the system when the newest version will run it fine. It kinda like they are contradicting themselves. First it wasn't stable and now it is and i never changed anything except the version #of the app?
Anyways i apologize for trapping you into this ordeal. To confirm the above screenshot and all what i said about the 2 versions are absolutely true at least in my case.
 
I don't doubt that you're telling the truth :)

Sorry if I sounded mean, or offended you in any way.

It's just that almost 100% of the time on here, people try prime and say: "Oh look, I can run Quake fine... that means that even though I fail prime, I can still do everything else, even with this massive overclock..."

Personally, I test prime STOCK before I start overclocking. If it can't run at stock, I do other burn-in tests. If they all pass (some are linux, so it can't be OS related), then I'll assume that prime has an error. It sounds like you did about that. Most don't.

IF, however, Prime runs fine at stock speeds, and fails when overclocked, then you're pwned by overclocking. Then we can make the assumptions set above. Sound like a good compromise?

:)
 
prime95 doesn't matter one click of a (ball) mouse if everything else you use your comp. for is stable.
 
You must log in or register to reply here.
Back
Top