Upgrade woes, Error starting Core post GPU upgrade

[Voyager]

I upgraded from a GeForce 8800 GTX to a GeForce GTX 480, and now theFolding@home core is failing to start. I'm not even entirely sure where to look for the problem.

OS: Windows 7 64 bit
CPU: Core 2 Duo E8400
8GB Ram

Was Geforce 8800 GTX running driver version 196.21
Is GeForce GTX 480 running driver version 258.96

Last entry in the log file is:

Code: Select all

[01:45:31] *------------------------------*
[01:45:31] Folding@Home GPU Core
[01:45:31] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[01:45:31]
[01:45:31] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[01:45:31] Build host: amoeba
[01:45:31] Board Type: Nvidia
[01:45:31] Core      :
[01:45:31] Preparing to commence simulation
[01:45:31] - Ensuring status. Please wait.
[01:45:40] - Looking at optimizations...
[01:45:40] - Working with standard loops on this execution.
[01:45:40] Examination of work files indicates 8 consecutive improper terminations of core.
[01:45:40] - Expanded 62948 -> 336763 (decompressed 534.9 percent)
[01:45:40] Called DecompressByteArray: compressed_data_size=62948 data_size=336763, decompressed_data_size=336763 diff=0
[01:45:40] - Digital signature verified
[01:45:40]
[01:45:40] Project: 10501 (Run 70, Clone 1, Gen 261)
[01:45:40]
[01:45:40] Entering M.D.
[01:45:46] Will resume from checkpoint file
[01:45:46] Tpr hash work/wudata_06.tpr:  1452820340 2432133128 2159739637 977467539 202991721
[01:45:46]
[01:45:46] Calling fah_main args: 14 usage=100
[01:45:46]
[01:45:49] CoreStatus = 63 (99)
[01:45:49] + Error starting Folding@Home core.
[01:45:49]
Folding@Home will go to sleep for 1 day as there have been 5 consecutive Cores executed which failed to complete a work unit.
[01:45:49] (To wake it up early, quit the application and restart it.)
[01:45:49] If problems persist, please visit our website at http://folding.stanford.edu for help.
[01:45:49] + Sleeping...


I have attempted reinstalling it, but that did not resolve the problem. I also noticed that it retained all of my user specific information, which leads me to suspect that it has retained part of the auto-config for the G80 card, and it I can get it to purge that, then it will fix itself. Unfortunately, I don't know where it stores then, and I have to leave fairly soon.

Any one have any ideas, or requests for additional data?

Harry Voyager

[Tobit]

It looks like you are still running the GPU2 client. Fermi cards cannot crunch GPU2 WUs which you were assigned. To use your new Fermi card, you need to upgrade to the GPU3 client.

[PantherX]

Welcome to the F@H Forum Harry,

I believe that you are using the Systray Version, If that is true, then:

For Systray Version:
Step A: Press Windows Key + R to open Run window
Step B: Type this %APPDATA% (with the % sign) and press enter
Step C: A window will appear (C:\Users\Username\AppData\Roaming) and there you have to delete a folder called "Folding@Home GPU"
Step D: Delete any shortcuts you have made
Note: Make sure that you delete the FahCore_xx, queue.dat and work folder as incorrect version\type will cause problems and won't allow your F@H GPU Client to function normally

Hopefully it will solve your problem.

[Voyager]

That makes sense. Is there a new client version I download, or will it upgrade automatically once my WU have expired? I checked around on the downloads page, and the most recent one listed was version 6.23, which I believe is what I've already got installed.

I did notice that the EVGA installer menu had a link to an EVGA folding@home application, and I would think they would include the right version with their cards, so I'll try that next.

Is there a way to release the GPU2 work units I've already got, rather than waiting for them to expire naturally? Seems like a waste to sit on them.

Thank you,

Harry Voyager

[PantherX]

For your GTX 460 you will need GPU3 BETA Client v6.32 that is on the this page: http://folding.stanford.edu/English/DownloadWinOther

The Client will not upgrade automatically.

Use the -oneunit flag on the GPU2 Client to avoid loosing the WU, method: http://fahwiki.net/index.php/How_do_I_a ... _client%3F

[Voyager]

Thank you, the GPU3 beta client worked. I think my ns/day jumped by a factor of 10.

[PantherX]

...I think my ns/day jumped by a factor of 10.

If you are using the Viewer, pleae avoid it as it is known to be buggy/resource hog. It will slow down the WU processing.

[BigApplePi]

New GTX 460 card fails to start new WU. Console version of GUI. I used the driver that came with the card and deleted the old ATI driver. An ideas?

[21:18:59] Entering M.D.
[21:19:05] Tpr hash work/wudata_01.tpr: 3585207269 2033162963 4018001788 4235377605 2145143473
[21:19:05]
[21:19:05] Calling fah_main args: 14 usage=100
[21:19:05]
[21:19:09] CoreStatus = 63 (99)
[21:19:09] + Error starting Folding@Home core.
[21:19:09] - Attempting to download new core...
[21:19:09] + Downloading new core: FahCore_11.exe
[21:19:09] + 10240 bytes downloaded
[21:19:09] + 20480 bytes downloaded
-
-
-
[21:19:11] + 665067 bytes downloaded
[21:19:11] Verifying core Core_11.fah...
[21:19:11] Signature is VALID
[21:19:11]
[21:19:11] Trying to unzip core FahCore_11.exe
[21:19:11] Decompressed FahCore_11.exe (1908736 bytes) successfully
[21:19:16] + Core successfully engaged
[21:19:21]
[21:19:21] + Processing work unit
[21:19:21] Core required: FahCore_11.exe
[21:19:21] Core found.
[21:19:21] Working on queue slot 01 [September 25 21:19:21 UTC]
[21:19:21] + Working ...
[21:19:21]
[21:19:21] *------------------------------*
[21:19:21] Folding@Home GPU Core
[21:19:21] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[21:19:21]
[21:19:21] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[21:19:21] Build host: amoeba
[21:19:21] Board Type: Nvidia
[21:19:21] Core :
[21:19:21] Preparing to commence simulation
[21:19:21] - Ensuring status. Please wait.
[21:19:31] - Looking at optimizations...
[21:19:31] - Working with standard loops on this execution.
[21:19:31] Examination of work files indicates 8 consecutive improper terminations of core.
[21:19:31] - Expanded 73833 -> 383588 (decompressed 519.5 percent)
[21:19:31] Called DecompressByteArray: compressed_data_size=73833 data_size=383588, decompressed_data_size=383588 diff=0
[21:19:31] - Digital signature verified
[21:19:31]
[21:19:31] Project: 6600 (Run 2, Clone 388, Gen 243)
[21:19:31]
[21:19:31] Entering M.D.
[21:19:37] Tpr hash work/wudata_01.tpr: 3585207269 2033162963 4018001788 4235377605 2145143473
[21:19:37]
[21:19:37] Calling fah_main args: 14 usage=100
[21:19:37]
[21:19:39] CoreStatus = 63 (99)
[21:19:39] + Error starting Folding@Home core.
[21:19:44]
[21:19:44] + Processing work unit
[21:19:44] Core required: FahCore_11.exe
[21:19:44] Core found.
[21:19:45] Working on queue slot 01 [September 25 21:19:45 UTC]
[21:19:45] + Working ...
[21:19:45]
[21:19:45] *------------------------------*
[21:19:45] Folding@Home GPU Core
[21:19:45] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[21:19:45]
[21:19:45] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[21:19:45] Build host: amoeba
[21:19:45] Board Type: Nvidia
[21:19:45] Core :
[21:19:45] Preparing to commence simulation
[21:19:45] - Ensuring status. Please wait.
[21:19:54] - Looking at optimizations...
[21:19:54] - Working with standard loops on this execution.
[21:19:54] Examination of work files indicates 8 consecutive improper terminations of core.
[21:19:54] - Expanded 73833 -> 383588 (decompressed 519.5 percent)
[21:19:54] Called DecompressByteArray: compressed_data_size=73833 data_size=383588, decompressed_data_size=383588 diff=0
[21:19:54] - Digital signature verified
[21:19:54]
[21:19:54] Project: 6600 (Run 2, Clone 388, Gen 243)
[21:19:54]
[21:19:54] Entering M.D.
[21:20:00] Tpr hash work/wudata_01.tpr: 3585207269 2033162963 4018001788 4235377605 2145143473
[21:20:00]
[21:20:00] Calling fah_main args: 14 usage=100
[21:20:00]
[21:20:05] CoreStatus = 63 (99)
[21:20:05] + Error starting Folding@Home core.
[21:20:10]
[21:20:10] + Processing work unit
[21:20:10] Core required: FahCore_11.exe
[21:20:10] Core found.
[21:20:10] Working on queue slot 01 [September 25 21:20:10 UTC]
[21:20:10] + Working ...
[21:20:10]
[21:20:10] *------------------------------*
[21:20:10] Folding@Home GPU Core
[21:20:10] Version 1.31 (Tue Sep 15 10:57:42 PDT 2009)
[21:20:10]
[21:20:10] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[21:20:10] Build host: amoeba
[21:20:10] Board Type: Nvidia
[21:20:10] Core :
[21:20:10] Preparing to commence simulation
[21:20:10] - Ensuring status. Please wait.
[21:20:19] - Looking at optimizations...
[21:20:19] - Working with standard loops on this execution.
[21:20:19] Examination of work files indicates 8 consecutive improper terminations of core.
[21:20:19] - Expanded 73833 -> 383588 (decompressed 519.5 percent)
[21:20:19] Called DecompressByteArray: compressed_data_size=73833 data_size=383588, decompressed_data_size=383588 diff=0
[21:20:19] - Digital signature verified
[21:20:19]
[21:20:19] Project: 6600 (Run 2, Clone 388, Gen 243)
[21:20:19]
[21:20:19] Entering M.D.
[21:20:25] Tpr hash work/wudata_01.tpr: 3585207269 2033162963 4018001788 4235377605 2145143473
[21:20:25]
[21:20:25] Calling fah_main args: 14 usage=100
[21:20:25]
[21:20:28] CoreStatus = 63 (99)
[21:20:28] + Error starting Folding@Home core.
[21:20:28]
Folding@Home will go to sleep for 1 day as there have been 5 consecutive Cores executed which failed to complete a work unit.
[21:20:28] (To wake it up early, quit the application and restart it.)
[21:20:28] If problems persist, please visit our website at http://folding.stanford.edu for help.
[21:20:28] + Sleeping...

[bruce]

There are three distinct types of GPUs: ATI, NVidia_g80, and NVidia_Fermi. When you change from one to another you must upgrade both the drivers and the FAH client. The portion of the log that you show seems to indicate you're running a project designed for the g80 using a core designed for the g80 which cannot be processed by Fermi.

I recommend you uninstall the client and delete the data files created in %APPDATA%. Install the new version of the client that supports Fermi and you'll probably be fine. For more information, read the GPU install guide. viewtopic.php?f=59&t=15196