Overclock.net banner

Ratio of Time Regarding to Stress Testing Apps

448 Views 8 Replies 5 Participants Last post by  suphiceto
1st, I am a firm believer when it comes to the more stress testing apps your OC can handle, the more stable it is. As I'm currently typing this, I'v just went beyond 8hrs of Prime95 (small fft). I'm trying to decide if I should stop @ the 12th hour or proceed to 24hrs. I will then test with large fft and then blend (same amount of time like small fft).

I need some opinions or possible facts on how long should I stress test with OCCT Linpack. I was thinking about 6 to 8hrs or simply set it for infinity & see how far it will last. We can all understand that it stress test the cpu a little harder but I think not by much. Any comments regarding to how long should I use OCCT?

As for LinX, I'm thinking 100 passes would suffice. Anybody disagree? Should I go for 150 or even 200 passes?

Usually, I would do 3hrs of OCCT, 6hrs of Prime95 (for all 3 parts), and 30 passes of LinX. I though for the 1st time, I will go the distance to really feel comfortable enough to say this OC is really stable.

All advice & comments will greatly be appreciated


***I use this rig for gaming, autocad, Photoshop, 3d Studio Max, & Accurender***
See less See more
1 - 9 of 9 Posts
Give it [email protected] for 2 WU's. If it doesn't crash, it's stable.
I've had numerous systems fail prime95 between 24 and 48 hours, as well as between 300 and 500 runs of LinX. The same adjustments that made such long runs stable in these stress tests also cleared up some very intermittent and mysterious problems elsewhere.

Anyway, for most people more than 24 hours of Prime, or more than 100 runs of LinX is usually redundant. However, I think 6 hours of prime and 30 passes of LinX have the opposite problem; there are some fairly serious instabilities that can remain hidden after such short tests.
Well, I usually go for 12 hours of prime. Since I NEVER have my PC that stressed for that long, I figure that is a good bench. I have seen errors after 2 hrs, 6hrs and 8hrs. You can even get them after 12 hours. the longer the betetr, but after a certain point, it is bragging rights more than anytihng esle.
Well, I guess I've been lucky because I've passed the shorter test procedures that I've mentioned with flying colors but I really do not feel comfortable to say it's stable. Honestly after further researching, I think it's very premature to say. I'm not really into folding (dont's see what I'm benefitting from that). I probably just need to take some time to read more about it.

So do you think it's safe to say after 12hrs of Prime95 (all parts) are enough? If you would contiue to stress test with OCCT, how long do you think would be considered a nice run when compared to 12hrs of Prime95 or even 24hrs?

I play games & use autocad on a daily basis. I usually use Photoshop & 3d Studio Max about several times within a month.
Quote:


Originally Posted by Blameless
View Post

I've had numerous systems fail prime95 between 24 and 48 hours, as well as between 300 and 500 runs of LinX. The same adjustments that made such long runs stable in these stress tests also cleared up some very intermittent and mysterious problems elsewhere.

Anyway, for most people more than 24 hours of Prime, or more than 100 runs of LinX is usually redundant. However, I think 6 hours of prime and 30 passes of LinX have the opposite problem; there are some fairly serious instabilities that can remain hidden after such short tests.

So do you feel that it's not necessary to deal with OCCT? Prime95 & LinX are respected more regarding to stress testing? Any negatives of OCCT I should be aware of?
See less See more
[email protected] is trying to find the cure to various protein-related diseases. Basically your CPU will simulate how proteins fold. It's great for stress testing, I've found problems in 10 minutes that Prime95 couldn't in 6 hours.
Quote:


Originally Posted by drBlahMan
View Post

So do you feel that it's not necessary to deal with OCCT? Prime95 & LinX are respected more regarding to stress testing? Any negatives of OCCT I should be aware of?

I use far more than Prime95 and LinX, and have no problem with OCCT. However, I do find it a bit redundant if you are already testing with the former. Nothing really wrong with that though.

Combination testing is also good, at times. After I've confirmed stability with various single tests, I'll use multiple ones to hit every subsystem possible, as hard as possible. Prime95 Small FFTs + HCI Memtest + looped HDD benchmarks + FurMark, all while watching a HD movie, for example. After all, real world programs and tasks can easily make demands of multiple components. I often do some fairly heavy duty encoding while gaming.
See less See more
You might wanna read this. (Read especially TwoCable's Comments)
1 - 9 of 9 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top