Problem with F@H GPU client

[tontod]

I recently upgraded to a 9800GT. It processes a WU fine, but I get this upon completion:


[04:01:22] Completed 100%
[04:01:22] Successful run
[04:01:22] DynamicWrapper: Finished Work Unit: sleep=10000
[04:01:32] Reserved 146176 bytes for xtc file; Cosm status=0
[04:01:32] Allocated 146176 bytes for xtc file
[04:01:32] - Reading up to 146176 from "work/wudata_07.xtc": Read 146176
[04:01:32] Read 146176 bytes from xtc file; available packet space=786284288
[04:01:32] xtc file hash check passed.
[04:01:32] Reserved 22248 22248 786284288 bytes for arc file=<work/wudata_07.trr> Cosm status=0
[04:01:32] Allocated 22248 bytes for arc file
[04:01:32] - Reading up to 22248 from "work/wudata_07.trr": Read 22248
[04:01:32] Read 22248 bytes from arc file; available packet space=786262040
[04:01:32] trr file hash check passed.
[04:01:32] Allocated 560 bytes for edr file
[04:01:32] Read bedfile
[04:01:32] edr file hash check passed.
[04:01:32] Logfile not read.
[04:01:32] GuardedRun: success in DynamicWrapper
[04:01:32] GuardedRun: done
[04:01:32] Run: GuardedRun completed.
[04:01:34] + Opened results file
[04:01:34] - Writing 169496 bytes of core data to disk...
[04:01:34] Done: 168984 -> 167536 (compressed to 99.1 percent)
[04:01:34] ... Done.
[04:01:34] DeleteFrameFiles: successfully deleted file=work/wudata_07.ckp
[04:01:34] Shutting down core
[04:01:34]
[04:01:34] Folding@home Core Shutdown: FINISHED_UNIT
[04:01:36] CoreStatus = 64 (100)
[04:01:36] Sending work to server
[04:01:36] Project: 5781 (Run 0, Clone 693, Gen 2)


[04:01:36] + Attempting to send results [January 31 04:01:36 UTC]
[04:01:38] + Results successfully sent
[04:01:38] Thank you for your contribution to Folding@Home.
[04:01:38] + Number of Units Completed: 2

[04:01:42] - Preparing to get new work unit...
[04:01:42] + Attempting to get work packet
[04:01:42] - Connecting to assignment server
[04:01:42] - Successful: assigned to (171.67.108.11).
[04:01:42] + News From Folding@Home: Welcome to Folding@Home
[04:01:43] Loaded queue successfully.
[04:01:44] + Closed connections
[04:01:44]
[04:01:44] + Processing work unit
[04:01:44] Core required: FahCore_11.exe
[04:01:44] Core found.
[04:01:44] Working on queue slot 08 [January 31 04:01:44 UTC]
[04:01:44] + Working ...
[04:01:44]
[04:01:44] *------------------------------*
[04:01:44] Folding@Home GPU Core
[04:01:44] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[04:01:44]
[04:01:44] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[04:01:44] Build host: amoeba
[04:01:44] Board Type: Nvidia
[04:01:44] Core :
[04:01:44] Preparing to commence simulation
[04:01:44] - Looking at optimizations...
[04:01:44] DeleteFrameFiles: successfully deleted file=work/wudata_08.ckp
[04:01:44] - Created dyn
[04:01:44] - Files status OK
[04:01:44] - Expanded 46656 -> 252912 (decompressed 542.0 percent)
[04:01:44] Called DecompressByteArray: compressed_data_size=46656 data_size=252912, decompressed_data_size=252912 diff=0
[04:01:44] - Digital signature verified
[04:01:44]
[04:01:44] Project: 5765 (Run 12, Clone 162, Gen 1598)
[04:01:44]
[04:01:44] Assembly optimizations on if available.
[04:01:44] Entering M.D.
[04:01:50] Tpr hash work/wudata_08.tpr: 3805432376 217075550 1361196462 3786111867 1484395310
[04:01:50]
[04:01:50] Calling fah_main args: 14 usage=100
[04:01:50]
[04:01:50] Working on Protein
[04:01:51] Client config found, loading data.
[04:01:51] Starting GUI Server
[04:01:51] mdrun_gpu returned
[04:01:51] NANs detected on GPU
[04:01:51]
[04:01:51] Folding@home Core Shutdown: UNSTABLE_MACHINE


Not sure where the UNSTABLE_MACHINE error is coming from - I havent overclocked the GPU. It runs fine if I shut down the client, delete the work folder, then restart the client. Any ideas as to what could be going on?

 

[theAnimal]

Since you're not OCed, it's probably a bad WU.

 

[tontod]

Since you're not OCed, it's probably a bad WU.

It happened twice in a row. Didnt happen for this last WU, I unchecked the box "Do NOT lock cores to specific CPU"

Its amazing what a new video card can do. I upgraded from a HD 3450 to 9800GT and its at least 10x faster

 

[GLeeM]

It happened twice in a row.

They sometimes send the same WU three times in a row if it errors out, (I guess in case it got garbled in the sending?)

Let us know if it keeps happening. Might be able to help.