Strange number of errors on different comps
23 Sep 2009 15:17:22 UTC
Topic 194546
(moderation:
I am having a strange number of client errors on different comps, occuring at about the same time. On one of them I changed overclocking settings, but on the other one I did not change a thing.
All other projects are running fine, torture tests with Prime95 are passing. This makes me wonder if my 3D and graphics settings for Einstein may have caused unexpected problems? I just did reset them to default, but I would prefer to look into this before risking more errors.
The comps are available via my profile, if you want to look into details. Thank you for you help in advance.
Strange number of errors on different comps
)
I've looked at a number of comp errors for both your hosts. In my opinion the most likely cause is the overclocking - for both hosts.
I run a large number of similar hosts and mine are overclocked as well. I've been running EAH for several years and I've seen many examples of similar problems. In virtually all cases, the problems have been completely cured by 3 particular checks/actions:-
* make sure you are not inadvertently overclocking your RAM too much
* if the first two are OK, reduce your CPU overclock until you achieve stability
The apps of other projects don't seem to be as hard on the machine as the EAH app and I've also seen supposed prime stability give errors like yours - are you stressing all 4 cores on your quad core? How long are you letting the stress test run for? Just bite the bullet and back off on the overclock.
Cheers,
Gary.
Thank you, Gary. Ironically
)
Thank you, Gary. Ironically the old Athlon is overclocked since I got it in 2005 and contributed the bulk of my EAH credits. OK, other apps those days, so this does not have to tell much. It won't be too hard to keep it out since it's just a backup now and can occasionally crunch for any other project when it's running.
Regarding the Quad: I did run the stress test at least 4 times on all cores and once for more than half an hour, maybe not long enough. I witnessed other problems as well lately and now I am trying to reduce the undervolting. Last effort to save the current settings. Hopefully the next EAH runs will show the problems are solved now.