P7623 from 171.64.65.105 should not be assigned to pre-Fermi

[rhavern]

I have two 8800 GTS 512M (G92 based) GPUs that have been running for a couple of years. Last night they were assigned FahCore_15.exe WU, which if I understand correctly are GPU3 WU and require a Fermi or newer to process. Unsurprisingly, these GPUs hung. Here's the log, all retained if more is required. I've only included one, as the other looks much the same, just a different RCG of the same project. Nothing has changed on this box in months/years. Note that this box has 2x 8800 GTS 512M and one GTX 460 and has done for months/years without issue.

W7-32 SP1 all patches up to date, dedicated folder, nvidia drivers 285.62. Also observe that the GPU type and species is correctly identified in the log.

Code: Select all

# Windows GPU Systray Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.41r2

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: c:\gpu2
Arguments: -verbosity 9 -gpu 1 -forcegpu nvidia_g80 -<beta flag elided>
<snip>
[22:13:32] Initial: 0000; - Uploaded at ~34 kB/s
[22:13:32] - Averaged speed for that direction ~48 kB/s
[22:13:32] + Results successfully sent
[22:13:32] Thank you for your contribution to Folding@Home.
[22:13:32] + Number of Units Completed: 3884

[22:13:36] Trying to send all finished work units
[22:13:36] + No unsent completed units remaining.
[22:13:36] - Preparing to get new work unit...
[22:13:36] Cleaning up work directory
[22:13:36] + Attempting to get work packet
[22:13:36] Passkey found
[22:13:36] - Will indicate memory of 1023 MB
[22:13:36] Gpu type=2 species=11.
[22:13:36] - Connecting to assignment server
[22:13:36] Connecting to http://assign-GPU.stanford.edu:8080/
[22:13:37] Posted data.
[22:13:37] Initial: 40AB; - Successful: assigned to (171.64.65.105).
[22:13:37] + News From Folding@Home: Welcome to Folding@Home
[22:13:37] Loaded queue successfully.
[22:13:37] Gpu type=2 species=11.
[22:13:37] Sent data
[22:13:37] Connecting to http://171.64.65.105:8080/
[22:13:38] Posted data.
[22:13:38] Initial: 0000; - Receiving payload (expected size: 125036)
[22:13:39] - Downloaded at ~122 kB/s
[22:13:39] - Averaged speed for that direction ~56 kB/s
[22:13:39] + Received work.
[22:13:39] Trying to send all finished work units
[22:13:39] + No unsent completed units remaining.
[22:13:39] + Closed connections
[22:13:39] 
[22:13:39] + Processing work unit
[22:13:39] Core required: FahCore_15.exe
[22:13:39] Core not found.
[22:13:39] - Core is not present or corrupted.
[22:13:39] - Attempting to download new core...
[22:13:39] + Downloading new core: FahCore_15.exe
[22:13:39] Downloading core (/~pande/Win32/x86/NVIDIA/Fermi/Core_15.fah from www.stanford.edu)
[22:13:39] Initial: AFDE; + 10240 bytes downloaded
<snip>
[22:13:47] Initial: 2D59; + 1559166 bytes downloaded
[22:13:47] Verifying core Core_15.fah...
[22:13:47] Signature is VALID
[22:13:47] 
[22:13:47] Trying to unzip core FahCore_15.exe
[22:13:53] Decompressed FahCore_15.exe (4685824 bytes) successfully
[22:13:58] + Core successfully engaged
[22:14:03] 
[22:14:03] + Processing work unit
[22:14:03] Core required: FahCore_15.exe
[22:14:03] Core found.
[22:14:03] Working on queue slot 04 [August 6 22:14:03 UTC]
[22:14:03] + Working ...
[22:14:03] - Calling '.\FahCore_15.exe -dir work/ -suffix 04 -nice 19 -priority 96 -checkpoint 15 -verbose -lifeline 1744 -version 641'

[22:14:03] 
[22:14:03] *------------------------------*
[22:14:03] Folding@Home GPU Core
[22:14:03] Version                2.22 (Thu Dec 8 17:08:05 PST 2011)
[22:14:03] Build host             SimbiosNvdWin7
[22:14:03] Board Type             NVIDIA/CUDA
[22:14:03] Core                   15
[22:14:03] 
[22:14:03] Window's signal control handler registered.
[22:14:03] Preparing to commence simulation
[22:14:03] - Looking at optimizations...
[22:14:03] DeleteFrameFiles: successfully deleted file=work/wudata_04.ckp
[22:14:03] - Created dyn
[22:14:03] - Files status OK
[22:14:03] sizeof(CORE_PACKET_HDR) = 512 file=<>
[22:14:04] - Expanded 124524 -> 501826 (decompressed 402.9 percent)
[22:14:04] Called DecompressByteArray: compressed_data_size=124524 data_size=501826, decompressed_data_size=501826 diff=0
[22:14:04] - Digital signature verified
[22:14:04] 
[22:14:04] Project: 7623 (Run 41, Clone 1, Gen 11)
[22:14:04] 
[22:14:04] Assembly optimizations on if available.
[22:14:04] Entering M.D.
[22:14:06] Tpr hash work/wudata_04.tpr:  1825332569 4091132314 3306796877 623270375 3957477434
[22:14:06] GPU device info: vendor=0 device=0 name=<NA> match=0
[22:14:06] Working on Protein
[22:14:06] Client config found, loading data.
[22:14:06] Starting GUI Server
[22:35:47] - Autosending finished units... [August 6 22:35:47 UTC]
[22:35:47] Trying to send all finished work units
[22:35:47] + No unsent completed units remaining.
[22:35:47] - Autosend completed
[22:35:47] + Working...
[04:35:47] - Autosending finished units... [August 7 04:35:47 UTC]
[04:35:48] Trying to send all finished work units
[04:35:48] + No unsent completed units remaining.
[04:35:48] - Autosend completed
[04:35:48] + Working...
[07:09:43] ***** Got a SIGTERM signal (2)
[07:09:43] Killing all core threads

Folding@Home Client Shutdown.

EDIT: Deleted WU and queue.dat from one pre-Fermi GPU, was assigned a WU from same project/server.

[dnamechanic]

Similar situation here.

Two systems with 8800GT using client 6.23 for months have been assigned P7623 which they cannot fold.

Delete files, and server reassigns P7623 from same project/server.

[noprob]

Similar situation here.

Two systems with 8800GT using client 6.23 for months have been assigned P7623 which they cannot fold.

Delete files, and server reassigns P7623 from same project/server.

Why are you not using the updated console client?

as posted in the previous post before yours,may help.

I've also posted here with two Alpha Dawgs 8800GS gpu's 384MB folding just fine though one of the gpu's did have a problem for 5 hours and went to sleep waking to download yet another core15 WU to succeed in working correctly according to the log file.

[jcoffland]

This was caused by a recent assignment server code change. I believe it is now fixed.

[rhavern]

This is now resolved, thanks Joe.

EDIT: I had to delete the existing WU in order to proceed.