![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
So, I'm one of those paranoid people who likes to be certain that my system is 110% stable. Yes, I realize that some don't mind that your OCs cause a Prime95 error after 8 hours of Prime95 or whatever. I'm just one of those a-retentive people who like to know that my computer can do whatever it is want it to do indefinitely without errors. So, I normally run a 24-hour stress test on a new system at stock and then at OCed speed.
Well. . . here's the kicker. . . I overslept past the time I was supposed to stop my 24 hour dual-Prime95 and 3DMark06 (1920x1200 8aa/16af) loop. And 27 hours into it, 3DMark06 crashed. . . though somewhat gracefully. . . Prime95(s) kept running and the computer just threw up the "This application has generated an error and most be closed. . . send report/don't send".
To me, the idea behind a 24-hour test is that if it can make 24 hours, it's probably not going to ever error. So, seeing an error of some kind at 27 hours renders that assumption moot.
Now, being the type of guy who always gets neurotic and paranoid (about stability and possible problems) while building my own system (yet I often tell people not to worry about obscure stuff when helping friends/family). . . I'm now of course wishing that I had woken up and stopped the loop(s) at 24 hours. What I don't know can't hurt me!
Anyways. . . if you were a betting person, would you chalk the application error up to:
1. Bad code in new release of 3DMark06 (1.1.0.0). Application issue.
2. Bad driver from NVIDIA. Driver issue.
3. Unstable hardware. Hardware issue.
I suspect that I'll just call it stable and stop testing (dying to actually play on this system rather than stare at 24-hour stability tests). But, it just sorta makes me uneasy when my system can't do what it's supposed to do indefinitely.
Well. . . here's the kicker. . . I overslept past the time I was supposed to stop my 24 hour dual-Prime95 and 3DMark06 (1920x1200 8aa/16af) loop. And 27 hours into it, 3DMark06 crashed. . . though somewhat gracefully. . . Prime95(s) kept running and the computer just threw up the "This application has generated an error and most be closed. . . send report/don't send".
To me, the idea behind a 24-hour test is that if it can make 24 hours, it's probably not going to ever error. So, seeing an error of some kind at 27 hours renders that assumption moot.
Now, being the type of guy who always gets neurotic and paranoid (about stability and possible problems) while building my own system (yet I often tell people not to worry about obscure stuff when helping friends/family). . . I'm now of course wishing that I had woken up and stopped the loop(s) at 24 hours. What I don't know can't hurt me!
Anyways. . . if you were a betting person, would you chalk the application error up to:
1. Bad code in new release of 3DMark06 (1.1.0.0). Application issue.
2. Bad driver from NVIDIA. Driver issue.
3. Unstable hardware. Hardware issue.
I suspect that I'll just call it stable and stop testing (dying to actually play on this system rather than stare at 24-hour stability tests). But, it just sorta makes me uneasy when my system can't do what it's supposed to do indefinitely.
