New Nvidia drivers - BEWARE

Snapshot

Snapshot

Snapshot

Soldato
Joined
11 Oct 2006
Posts
5,172
Location
Wiltshire
Just tried the 295.73 drivers and they caused my GTX570 to hang after three or four percent. There's also a memory leak so that the memory used by the client isn't released when it's stopped and after this has happened a few times the card's memory usage is at max. It also significantly slows the SMP client on the same PC. I've gone back to 285.58.
 
Just updated the other day & now getting this:(

[12:19:24] Examination of work files indicates 8 consecutive improper terminations of core.
[12:19:24] sizeof(CORE_PACKET_HDR) = 512 file=<>
[12:19:24] - Expanded 121370 -> 544418 (decompressed 448.5 percent)
[12:19:24] Called DecompressByteArray: compressed_data_size=121370 data_size=544418, decompressed_data_size=544418 diff=0
[12:19:24] - Digital signature verified
[12:19:24]
[12:19:24] Project: 8031 (Run 16, Clone 18, Gen 8)
[12:19:24]
[12:19:24] Entering M.D.
[12:19:26] Tpr hash work/wudata_08.tpr: 121833354 2394170768 1371763611 2281816488 3797604197
[12:19:26] calling fah_main gpuDeviceId=0
[12:19:26] Working on Protein
[12:19:26] Client config found, loading data.
[12:19:29] CoreStatus = 63 (99)
[12:19:29] + Error starting Folding@home core.
[12:19:34]
[12:19:34] + Processing work unit
[12:19:34] Core required: FahCore_15.exe
[12:19:34] Core found.
[12:19:34] Working on queue slot 08 [February 23 12:19:34 UTC]
[12:19:34] + Working ...
[12:19:34] - Calling '.\FahCore_15.exe -dir work/ -suffix 08 -nice 19 -checkpoint 15 -verbose -lifeline 3708 -version 641'

[12:19:34]
[12:19:34] *------------------------------*
[12:19:34] Folding@Home GPU Core
[12:19:34] Version 2.20 (Tue Aug 2 12:06:37 PDT 2011)
[12:19:34] Build host SimbiosNvdWin7
[12:19:34] Board Type NVIDIA/CUDA
[12:19:34] Core 15
[12:19:34]
[12:19:34] Window's signal control handler registered.
[12:19:34] Preparing to commence simulation
[12:19:34] - Ensuring status. Please wait.
[12:19:43] - Looking at optimizations...
[12:19:43] - Working with standard loops on this execution.
[12:19:43] Examination of work files indicates 8 consecutive improper terminations of core.
[12:19:43] sizeof(CORE_PACKET_HDR) = 512 file=<>
[12:19:43] - Expanded 121370 -> 544418 (decompressed 448.5 percent)
[12:19:43] Called DecompressByteArray: compressed_data_size=121370 data_size=544418, decompressed_data_size=544418 diff=0
[12:19:43] - Digital signature verified
[12:19:43]
[12:19:43] Project: 8031 (Run 16, Clone 18, Gen 8)
[12:19:43]
[12:19:43] Entering M.D.
[12:19:45] Tpr hash work/wudata_08.tpr: 121833354 2394170768 1371763611 2281816488 3797604197
[12:19:45] calling fah_main gpuDeviceId=0
[12:19:45] Working on Protein
[12:19:45] Client config found, loading data.
[12:19:48] CoreStatus = 63 (99)
[12:19:48] + Error starting Folding@home core.
[12:19:48] - Attempting to download new core...
[12:19:48] + Downloading new core: FahCore_15.exe
[12:19:48] Downloading core (/~pande/Win32/x86/NVIDIA/Fermi/Core_15.fah from www.stanford.edu)
Click to expand...
Will revert to old driver & see what happens, as it seems to crash after completion of each unit 8031 fermi. I have a 560.
 
I'm not sure what's going on now but it could be the card. I reverted the drivers, left it running while I was at work and it played silly wotsits all day. I created a new folder, machine ID, config, etc. and it's still being silly: 2:36 for the first seven frames, 16:44 for the eighth. I'm going to turn it off completely then run some benchies once it's cooled to see if it slows after a few minutes.
 
You must log in or register to reply here.
Back
Top Bottom