if a system passes prime95, memtest, and 3dmark...

is it 100% stable? or is there still a chance that under normal use, and heavy gaming errors can still arise?

I ask because I have my system at 200x13, it passes prime, and memtest(need to test long term still) but Im curious if theres a chance that even with these tests, things could still fail...

How long did you run Prime 95 for? Alot of poeple run the Torture test for over 8 hours for stability checking...if your still stable after that then you really have nothing to worry about! As the torture test runs memory and CPU at 100% I don't know any game that does that just yet!

Im not sure I understand prime...

how do you run the "torture test"?

I just turn it on and it runs. I havent found settings (other than setting it to high priority in alt-ctl-del). I didnt run it long, Im going to tonight, let it run all night.

Under the Options menu there's a "Torture Test" item. Also, if you want to ensure maximum stability you should also run 3dMark in the background on loop.

never saw the point of long stress tests, if ive changed some settings ill usualy doa 10min check to see if their is any 'major' instability, if its fine ill use the machine play games etc. if i see crashes etc then i know its unstable. cant see why anyone would wanna run stress tests for hours and hours on a system.

Stee said:

never saw the point of long stress tests, if ive changed some settings ill usualy doa 10min check to see if their is any 'major' instability, if its fine ill use the machine play games etc. if i see crashes etc then i know its unstable. cant see why anyone would wanna run stress tests for hours and hours on a system.

Click to expand...

after running stress tests for about a day total, you can be sure nothing like random crashes will ever occur(for the most part). your method is pretty much the same, but the testing takes even longer to produce pass/fail result.

stress testing also doesnt really harm the components unless they get too hot (which should be taken care of before then or be monitored during the test).

i dissagree and would say that stress testing is obviously gonna limit the components factor by a degree, maybe not huge, but with the rate some people change settings the stress test it i wouldnt be surprised if there was a drastic loss of life. as for my stress test taking me longer to find a instability, does that really matter? i can correct it when i get an error, meanwhile itll cause me no damage. thats my point why purposely go find an instability which is gonna make ur pc unusuable or performing badly due to it being busy constantly. why do that when u can let the instability find u and during that time u can play games etc.

never saw the point of long stress tests, if ive changed some settings ill usualy doa 10min check to see if their is any 'major' instability, if its fine ill use the machine play games etc. if i see crashes etc then i know its unstable. cant see why anyone would wanna run stress tests for hours and hours on a system.

Click to expand...

I'd rather run the tests overnight / over the weekend so that when I'm crunching that 100+ MB excel file I know nothing is going to jack up. Of course, I dont overclock; everything in my system is @ stock, but still, it's nice to know.

- Awperator

just run torture test number three on prime, and run something in the backround using the other 50 percent if you have HT.

Stee said:

which is gonna make ur pc unusuable or performing badly due to it being busy constantly.

Click to expand...

have you ever looked into folding (or seti for that matter)? these are distributed programs that run constantly and are supported by many of the forum memebers. my cpu is always at 100% usage. i have never heard of anyone say that their computer died because it was folding 24/7. if stress testing does damage components, it may lessen the life slightly, but by the time that would become a factor, you will have upgraded possibly multiple times

never said it would die... i said your limiting its life badly, and its not slight

I can't see it limting it's life badly. Running at 100% or running at 0%, it's still got power running through the transistors. Running 100% may increase the amount of heat created by the CPU, but my system @ full load is cooler than most OEM systems at idle Also, while the idle loop runs only 1 instruction, a system at 100% load will switch between many different instructions. Constantly sending HLT to a CPU would stress only whatever transistors are used to comunicate the "don't do anything right now" signal to the rest of the processor. At full load, the mix of instructions mean any single transistor used for a specific instruction will be stressed less than if it were the ONLY instruction used.

I've NEVER heard of a CPU dying from running it too much. I just can't immagine running a CPU at full load limiting it's life any more than would be useable.
JigPu

I don't remember the link to intel's website, but when they design chips, they are designed to run at load constantly. I want to say 10 years, but don't quote me on that. It is actually worse to have chips that the temps fluctuate a large amount. No difference than the roads in MN. :0

I would run Prime95 for as long as I could.. minimum three days. The more the better because it's less likely to generate errors after it has ran for 7 days fine.

Memtest86 I would run overnight, and you can also do the same with 3dmark as memtest runs in dos and 3dmark takes up the whole screen. 3dmark is probably a very good stress tester imho, it stresses out everything.

Stee said:

i dissagree and would say that stress testing is obviously gonna limit the components factor by a degree, maybe not huge, but with the rate some people change settings the stress test it i wouldnt be surprised if there was a drastic loss of life. as for my stress test taking me longer to find a instability, does that really matter? i can correct it when i get an error, meanwhile itll cause me no damage. thats my point why purposely go find an instability which is gonna make ur pc unusuable or performing badly due to it being busy constantly. why do that when u can let the instability find u and during that time u can play games etc.

Click to expand...

Well, when I was testing for overclocking, I would make sure the Torture test would run for 7 days straight without an error with the addition of other tests. Why? Because you keep your computer on for more than just a day and I want full stability when I'm running a computer. I've seen errors that generate after 72 hours and it keeps coming back.

I just can't stand an unstable computer. Knowing that whatever I do has a chance of some sort of error just makes me feel insecure on the computer I guess.

I think that the longer you run it, the less likely it will generate errors depending on how many days you run.. besides, the stress builds up and the longer ya run it, the better. People even do it long term with Folding or SETI.