View Full Version : Quicker stability tests? are there Any for Dual Core?
n91htmare
04-29-2006, 09:18 AM
I am overclocking my Opteron 165 but I dont want to wait 8 hours to find if the thing is stable, I mean i will do that once I find a stable setting.
I'm wondering if there is a quicker way to find stability. I know about OCCT and it's 30 min test, which is pretty good I think. But It doesnt seem to work on dual core and I can't open it twice to run two instances.
How would one go about this ?
Ic3man
04-29-2006, 09:20 AM
Run 2x32mb SuperPI.
There IS a way to run 2xocct but can't remember how, someone on here should know.
n91htmare
04-29-2006, 09:21 AM
Run 2x32mb SuperPI.
There IS a way to run 2xocct but can't remember how, someone on here should know.
I tried running two instances of superPI, But when i start one, the other gets an error... even on a system that was dual prime stable for 6 hours.
Maybe i'm not doing it right.
politenessman
04-29-2006, 09:28 AM
you are not doing it right. your system is fine. error is because both are trying to write to the same log file. has to be installed twice, to different folders, one named superpi1, the other superpi2( or whatever), start line has to have affinity. google it, step by step proc is around somewhere.
very truly yours,
politenessman
n91htmare
04-29-2006, 09:37 AM
you are not doing it right. your system is fine. error is because both are trying to write to the same log file. has to be installed twice, to different folders, one named superpi1, the other superpi2( or whatever), start line has to have affinity. google it, step by step proc is around somewhere.
very truly yours,
politenessman
Got it. Set affinity manually for now will write start line later.
THanks Polite.
I am overclocking my Opteron 165 but I dont want to wait 8 hours to find if the thing is stable, I mean i will do that once I find a stable setting.
I'm wondering if there is a quicker way to find stability. I know about OCCT and it's 30 min test, which is pretty good I think. But It doesnt seem to work on dual core and I can't open it twice to run two instances.
How would one go about this ?
Increase the clock and run prime for 5 minutes until it fails. Back down 50 MHz and run prime for 24 hours. The 50 MHz part may vary, but this is a quick way I use.
stummerwinter
04-29-2006, 12:10 PM
Try rosetta@home (http://www.xtremesystems.org/forums/showthread.php?t=84660)...;)
It came out for me, that Prime Custom > 10 h without error but r@h freezed...;)
It gaves you faster and better results...;)
Vapor
04-29-2006, 12:16 PM
Yup, R@H is the ultimate stability teest...but definitely not fast right at the edge of stability. One thing I've learned to like is priming on one core and OCCTing on the other, and then backing down ~25MHz from where that's stable.
Leon939
04-29-2006, 04:30 PM
I am overclocking my Opteron 165 but I dont want to wait 8 hours to find if the thing is stable, I mean i will do that once I find a stable setting.
I'm wondering if there is a quicker way to find stability. I know about OCCT and it's 30 min test, which is pretty good I think. But It doesnt seem to work on dual core and I can't open it twice to run two instances.
How would one go about this ?
try dual SP2004 or dual F@H :D
fareastgq
04-29-2006, 08:57 PM
fastest stability test for me = dual prime, dual super pi, 3dmark all at the same time, if the other progs are still running at the end of the 3dmark loop, you can bet your ass it's stable for 24. Of course, I play my fav game for 30 mins also after that. I don't have time to wait a damn day for it.
n91htmare
04-30-2006, 11:37 AM
Hey guys, I have a question and I Don't want to start a new thread.
I just primed my computer at 2898 (322x9) Small FFT.... and its still going strong for almost 9 hours.
BUT....
A Blend test will fail on core 1 after like 3 or so hours..... Could it be my memory? I have the 2GB HZ G.skill at 263 stock timings. (3-4-4-8)
What should I do to make more stabe....? more voltage to ram?
http://members.cox.net/rsxtypes/opty165oc3.JPG
HiJon89
04-30-2006, 11:41 AM
It could be a memory controller issue, the RAM shouldn't need more than 2.7-2.8v
creidiki
04-30-2006, 11:42 AM
Ive seen quite a lot of cases where LESS v makes them go better. Mine work fine @ 2.6v up to the point where they simply stop working no matter what I do.
stummerwinter
04-30-2006, 11:45 AM
What happens, when you use another divider?
Some failure?
SP2004 is less stress compared to Prime Custom...;)
n91htmare
04-30-2006, 11:59 AM
What happens, when you use another divider?
Some failure?
SP2004 is less stress compared to Prime Custom...;)
can you give me the prime custom values to use. I can run Dual Prime95.
I'm using 166 divider and I don't know how that will effect it.