one over half of my donate@home wus I get a computation error with an error message box that says test615.exe has a problem and to close or try to find a solution and then close. Is there a fix for it?
Printable View
one over half of my donate@home wus I get a computation error with an error message box that says test615.exe has a problem and to close or try to find a solution and then close. Is there a fix for it?
I've seen that, but not that frequently. What's your setup?
I have a Radeon HD 5700 series card. It runs collatz and moo wrapper just fine.Quote:
Originally Posted by Maxwell
Did you ever learn anything on this issue? I am having trouble on 2 of 3 PCs. I am was beginning to suspect it had something to do with dual AMD gpus and Donate. My Nividia dual is fine. But I see you are running one card and others have duals.Quote:
Originally Posted by artemis8
I am running the 12.3 drivers. More trouble with the 7970s but also getting issues sometimes on the 6970. Time to reinstall drivers.
no unfortunately. I've restarted boinc and rebooted my computer, but the error comes up a good percentage of the time and I have to keep an eye on it or it ties up the gpu until I tell it to close.
artemis8, if you referring to the Windows error window then disable Windows error reporting (Link).Quote:
Originally Posted by artemis8
Same here. This is the error Hydra6 is getting:Quote:
Originally Posted by artemis8
Faulting application name: test615.exe, version: 0.0.0.0, time stamp: 0x4f4779fa
Faulting module name: msvcrt.dll, version: 7.0.7601.17744, time stamp: 0x4eeaf722
Exception code: 0xc000000d
Fault offset: 0x0000987b
Faulting process id: 0x84c
Faulting application path: C:\ProgramData\BOINC\projects\donateathome.org\tes t615.exe
Faulting module path: C:\Windows\syswow64\msvcrt.dll
Testing out different ideas, I will let you know if I learn anything.
My intel/nividia setups no problems, my AMD/ATI setup are both having issues. On the other one I am just getting occasional computation errors (~10%). Wiping the driver install and restalling at least stopped the crashes.
Here is a thread covering the issue, looks like STE\/E has run into it also. There is no resolution in the thread, but it does run through a lot of what they tried.
http://dev.gpugrid.org/forum_thread.php?id=58&sort=6
I seem to be having luck on 2 fronts, we will see what happens overnight.
On the PC with application error (a.k.a. BEX error) I changed my Boinc preferences to limit max cpu to 75% (i.e. exclude one of the 4 cpus). This is the longest I have gone without the error. The error being thrown is a error on the CPU activity, so this was my guess on resolution. Time will tell.
On the PC with the computation errors, I suspended the CPU project to remove all CPU load except for support of Donate GPU activity. Not one computation error since and again the longest I have gone without that error on Donate. On this PC I suspect the North Bridge/Memory is the likely source of the problem. It was also the hardest part of the O.C. If I go error free over night, I will turn off O.C. settings and enable the CPU project again.
BTW running POEM on both for CPU project.
If I go that route, I think I will do it to just stop the error reporting on test615 (if I can) and not all of Windows (the link he provided mentions that possibility). Each time this error occurs, the WU still reports as valid. So, the error is basically just annoying and stops work.Quote:
Originally Posted by Fire$torm
Good tip Fire$torm, thanks.
Update.
The BEX error returned at 12:20 a.m. I disabled the CPU project on that PC also and no errors since. No computation errors on second PC since pulling the CPU project. I will let it continue to ths way for awhile longer.
So what are your CPUs doing? If under heavy load and your are still getting the error, try lightening the load and report back, please.
I don't called about the windows reporting error so much as the fact that it errors and I don't get any credit for the work unit.Quote:
Originally Posted by Fire$torm
I get this error on these wusQuote:
Originally Posted by artemis8
8 Apr 2012 | 3:27:15 UTC Error while computing 2,289.55 11.17 --- test alpha. Required: Internet connection, allow in antivirus, not for XP v6.15 (opencl_ati_100)
Update I posted on Donate Forum -
Last night on both systems I disabled their CPU project (POEM). The BEX error stopped and the computation errors stopped. Ran error free for over 12 hours.
Turned off all O.C. settings and turned on CPU projects again. On the Phenom x4 that gets the BEX error, I changed the CPU % to 75% (which only lets POEM run on 2 cpus because of the resource pulling of Donate). Eventually the BEX showed up again.
On the bulldozer. With CPU % at 100% and no O.C. I started to get computation errors again. I changed the CPU % to 75% (POEM gets 5 of the 8 CPUs). Its been 4 hours without Computation errors.
Whatever the issue is it seems to be related to the CPU activity. The more I reduce the utilization, the less freq. I get errors.
I am gong to stop doing Donate until after the primaboinca challenge.
OK. With Donate needing a full CPU core, I have set BOINC Manager on my quads to use only 75% of CPU resources. That has freed up enough for Donate to run alongside the CPU projects. Maybe give that a try?Quote:
Originally Posted by artemis8
I think I'll probably just pass on crunching Donate since I only have one computer to run it on and want to use my CPUs for crunching other things. Moo Wrapper will be getting my love (since collatz has been down for a while :( )
I like MilkyWay when I want max out of the CPUs. It seems real light, plus Moo and my systems don't play nice.Quote:
Originally Posted by artemis8
Don't know why I like MilkyWay so much, personality quirk of some sort.
Slicker said hopefully he should have it up sometime today if the problem is not bad hardware. He was on vacation and his dog didn't take to the server admin training...... :PQuote:
Originally Posted by artemis8