Prime95 always comes to mind for stress testing. If it passes, computer is stable, all is good. But what's a pass? How long? Which version?
I mean Prime95 is essentially a program that "tries" to crash your computer. What I would like to know is; can it make a stable computer crash? I'm wondering if anyone has ever looked at Prime95 in detail at a programming level to see if there are any artificial causes for crashes. For example, all it would potentially take to crash a computer is a slow memory leak which would eventually bog everything down and sooner or later trigger a crash. Now this is a bit of a chicken-before-the-egg kind of thing because if it doesn't pass prime, then it's not stable and therefore we don't really have a controlled computer that we know is "stabe". The best we can do is to assume that one @ stock frequencies/voltages is stable.
I propose a barrage of Pime95 tests on some of Tom's, or anyone elses, computers @ stock values to determine how many "stable" computers Prime95 can crash. Try different versions, different tests, look at things like total memory used to see if its slowl creeping up etc. I mean at the end of the day, i could probably spend 24 hours opening up 1000 copies of the same word program and at some point my computer would probably crash even though it's stable. What could this bring us? well maybe a more refined Prime95 program and maybe a few extra hertz for my overclock
Cheerios
I mean Prime95 is essentially a program that "tries" to crash your computer. What I would like to know is; can it make a stable computer crash? I'm wondering if anyone has ever looked at Prime95 in detail at a programming level to see if there are any artificial causes for crashes. For example, all it would potentially take to crash a computer is a slow memory leak which would eventually bog everything down and sooner or later trigger a crash. Now this is a bit of a chicken-before-the-egg kind of thing because if it doesn't pass prime, then it's not stable and therefore we don't really have a controlled computer that we know is "stabe". The best we can do is to assume that one @ stock frequencies/voltages is stable.
I propose a barrage of Pime95 tests on some of Tom's, or anyone elses, computers @ stock values to determine how many "stable" computers Prime95 can crash. Try different versions, different tests, look at things like total memory used to see if its slowl creeping up etc. I mean at the end of the day, i could probably spend 24 hours opening up 1000 copies of the same word program and at some point my computer would probably crash even though it's stable. What could this bring us? well maybe a more refined Prime95 program and maybe a few extra hertz for my overclock
Cheerios