F@H, videocrunching, EUE error

[Thump553]

Here's the situation: Win7 74 bit box, xfx GEForce GT 240, never overclocked and doesn't overheat has been getting a lot of EUE/UNSTABLE MACHINE errors in the last few days-almost every night. The machine was rebooted and I deleted the core to no effect. No hardware has been changed in months. All Windows updates current and otherwise the machine seems to be running fine.

I forgot to make a note of what CPU the machine has, but it is a basic Staples $300 Compaq desktop-the only change is the videocard and swapped out the PSU for an Antec 450 or 500 (forget), both done about a year ago.

Here's part of a log:

---------------
[09:53:15] *------------------------------*
[09:53:15] Folding@Home GPU Core
[09:53:15] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[09:53:15]
[09:53:15] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[09:53:15] Build host: amoeba
[09:53:15] Board Type: Nvidia
[09:53:15] Core :
[09:53:15] Preparing to commence simulation
[09:53:15] - Looking at optimizations...
[09:53:15] DeleteFrameFiles: successfully deleted file=work/wudata_02.ckp
[09:53:15] - Created dyn
[09:53:15] - Files status OK
[09:53:15] - Expanded 62848 -> 336799 (decompressed 535.8 percent)
[09:53:15] Called DecompressByteArray: compressed_data_size=62848 data_size=336799, decompressed_data_size=336799 diff=0
[09:53:15] - Digital signature verified
[09:53:15]
[09:53:15] Project: 10504 (Run 142, Clone 0, Gen 1236)
[09:53:15]
[09:53:15] Assembly optimizations on if available.
[09:53:15] Entering M.D.
[09:53:21] Tpr hash work/wudata_02.tpr: 3342642038 3113939638 3853827689 4155492032 1970456878
[09:53:21]
[09:53:21] Calling fah_main args: 14 usage=100
[09:53:21]
[09:53:21] mdrun_gpu returned
[09:53:21] Going to send back what have done -- stepsTotalG=0
[09:53:21] Work fraction=0.0000 steps=0.
[09:53:26] logfile size=4947 infoLength=4947 edr=0 trr=25
[09:53:26] + Opened results file
[09:53:26] - Writing 5485 bytes of core data to disk...
[09:53:26] Done: 4973 -> 1851 (compressed to 37.2 percent)
[09:53:26] ... Done.
[09:53:26] DeleteFrameFiles: successfully deleted file=work/wudata_02.ckp
[09:53:26]
[09:53:26] Folding@home Core Shutdown: UNSTABLE_MACHINE
[09:53:30] CoreStatus = 7A (122)
[09:53:30] Sending work to server
[09:53:30] Project: 10504 (Run 142, Clone 0, Gen 1236)
[09:53:30] - Read packet limit of 540015616... Set to 524286976.


[09:53:30] + Attempting to send results [April 10 09:53:30 UTC]
[09:53:30] + Results successfully sent
[09:53:30] Thank you for your contribution to Folding@Home.
[09:53:34] EUE limit exceeded. Pausing 24 hours.

Folding@Home Client Shutdown.

--------------------

Any idea what the problem is and what I should do to solve it?

 

[blckgrffn]

The fan on the GPU is spinning?

My vote? Your PSU is flaking out. I would try replacing it with a known good (best) or a new, decent quality unit and see if that resolves the issue.

 

[Thump553]

Fan is spinning. I do have a spare unused PSU around, I'll toss that in and see what happens.

 

[Thump553]

Replacing the PSU didn't fix the problem (replaced it with a brand new Antec eco 620). Still getting EUE errors now, fairly consistently.

Any other ideas on how to solve this problem?


BTW the computer is a Compaq CQ5320F (Athlon II X2 240, 2.8 GHz, 3 GB RAM), stock except for the PSU and videocard mentioned in the first post.

BTW #2: looking at the logs, it appears to process one workunit properly then choke on any subsequent ones.

 

[ZipSpeed]

Did you try running Memtest and Prime95 (or equivalent stress test)?

 

[Peter Trend]

My computer just started doing the exact same thing...

 

[Thump553]

Started running memtest this morning. By the time I left it had completed one complete cycle with no errors.

BTW temps in the case seem fine-62 degrees C was the highest reached by the GPU overnight, with a 39% highest gpu fan speed.

 

[Thump553]

Anyone have any other ideas? I uninstalled and reinstalled the latest version of Win32-gpu, running it as a console.

Now it seems to complete one workunit then go into EUE / unstable machine. If I shut down the console and restart it, it's fine-for one workunit.

 

[blckgrffn]

Sorry, I've got nothing Maybe try posting your issue - and what you've done to fix it - in the F@H forum?