I got the following info when this wu was attempted.
23:42:07] + Processing work unit
[23:42:07] Core required: FahCore_11.exe
[23:42:07] Core found.
[23:42:07] Working on queue slot 08 [July 20 23:42:07 UTC]
[23:42:07] + Working ...
[23:42:07] - Calling '.\FahCore_11.exe -dir work/ -suffix 08 -priority 96 -nocpulock -checkpoint 15 -verbose -lifeline 3220 -version 623'
[23:42:07]
[23:42:07] *------------------------------*
[23:42:07] Folding@Home GPU Core
[23:42:07] Version 1.27 (Thu Jun 18 14:02:10 PDT 2009)
[23:42:07]
[23:42:07] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[23:42:07] Build host: amoeba
[23:42:07] Board Type: Nvidia
[23:42:07] Core :
[23:42:07] Preparing to commence simulation
[23:42:07] - Looking at optimizations...
[23:42:07] DeleteFrameFiles: successfully deleted file=work/wudata_08.ckp
[23:42:07] - Created dyn
[23:42:07] - Files status OK
[23:42:07] - Expanded 46632 -> 252912 (decompressed 542.3 percent)
[23:42:07] Called DecompressByteArray: compressed_data_size=46632 data_size=252912, decompressed_data_size=252912 diff=0
[23:42:07] - Digital signature verified
[23:42:07]
[23:42:07] Project: 5765 (Run 0, Clone 320, Gen 155)
[23:42:07]
[23:42:07] Assembly optimizations on if available.
[23:42:07] Entering M.D.
[23:42:13] Tpr hash work/wudata_08.tpr: 3549813770 3971485031 2027333430 3964038508 2941887379
[23:42:13]
[23:42:13] Calling fah_main args: 14 usage=100
[23:42:13]
[23:42:13] Working on Protein
[23:42:14] Client config found, loading data.
[23:42:14] Starting GUI Server
[23:42:14] mdrun_gpu returned
[23:42:14] NANs detected on GPU
[23:42:14]
[23:42:14] Folding@home Core Shutdown: UNSTABLE_MACHINE
[23:42:17] CoreStatus = 7A (122)
[23:42:17] Sending work to server
[23:42:17] Project: 5765 (Run 0, Clone 320, Gen 155)
[23:42:17] - Read packet limit of 540015616... Set to 524286976.
[23:42:17] - Error: Could not get length of results file work/wuresults_08.dat
[23:42:17] - Error: Could not read unit 08 file. Removing from queue.
[23:42:17] Trying to send all finished work units
[23:42:17] + No unsent completed units remaining.
[23:42:17] - Preparing to get new work unit...
[23:42:17] + Attempting to get work packet
[23:42:17] - Will indicate memory of 8188 MB
[23:42:17] - Connecting to assignment server
[23:42:17] Connecting to http://assign-GPU.stanford.edu:8080/
[23:42:17] Posted data.
Then the following wu was fine.
Project: 5765 (Run 0, Clone 320, Gen 155)
Moderators: Site Moderators, FAHC Science Team