Folding issue

[Fuganater]

So my 560's are chipping away at WU's at 13-18k PPD until this happens...


[21:58:08] Completed 19000000 out of 50000000 steps (38%).
[21:58:30] CoreStatus = 63 (99)
[21:58:30] + Error starting Folding@home core.
[21:58:35]
[21:58:35] + Processing work unit
[21:58:35] Core required: FahCore_15.exe
[21:58:35] Core found.
[21:58:35] Working on queue slot 01 [June 13 21:58:35 UTC]
[21:58:35] + Working ...
[21:58:35]
[21:58:35] *------------------------------*
[21:58:35] Folding@Home GPU Core
[21:58:35] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[21:58:35]
[21:58:35] Build host: SimbiosNvdWin7
[21:58:35] Board Type: NVIDIA/CUDA
[21:58:35] Core : x=15
[21:58:35] Window's signal control handler registered.
[21:58:35] Preparing to commence simulation
[21:58:35] - Ensuring status. Please wait.
[21:58:44] - Looking at optimizations...
[21:58:44] - Working with standard loops on this execution.
[21:58:44] - Previous termination of core was improper.
[21:58:44] - Files status OK
[21:58:44] sizeof(CORE_PACKET_HDR) = 512 file=<>
[21:58:44] - Expanded 44228 -> 172159 (decompressed 389.2 percent)
[21:58:44] Called DecompressByteArray: compressed_data_size=44228 data_size=172159, decompressed_data_size=172159 diff=0
[21:58:44] - Digital signature verified
[21:58:44]
[21:58:44] Project: 6806 (Run 1369, Clone 2, Gen 42)
[21:58:44]
[21:58:44] Entering M.D.
[21:58:46] Will resume from checkpoint file work/wudata_01.ckp
[21:58:46] Tpr hash work/wudata_01.tpr: 2404883553 1113435923 1438457268 3050240718 1210848591
[21:58:46] Working on 2 PEPTIDE (1-42)
[21:58:46] Client config found, loading data.
[21:58:47] Resuming from checkpoint
[21:58:47] fcCheckPointResume: retreived and current tpr file hash:
[21:58:47] 0 2404883553 2404883553
[21:58:47] 1 1113435923 1113435923
[21:58:47] 2 1438457268 1438457268
[21:58:47] 3 3050240718 3050240718
[21:58:47] 4 1210848591 1210848591
[21:58:47] fcCheckPointResume: file hashes same.
[21:58:47] fcCheckPointResume: state restored.
[21:58:47] fcCheckPointResume: name work/wudata_01.log Verified work/wudata_01.log
[21:58:47] fcCheckPointResume: name work/wudata_01.trr Verified work/wudata_01.trr
[21:58:47] fcCheckPointResume: name work/wudata_01.xtc Verified work/wudata_01.xtc
[21:58:47] fcCheckPointResume: name work/wudata_01.edr Verified work/wudata_01.edr
[21:58:47] fcCheckPointResume: state restored 2
[21:58:47] Resumed from checkpoint
[21:58:47] Setting checkpoint frequency: 500000
[21:58:47] Completed 19000001 out of 50000000 steps (38%).
[21:58:47] Starting GUI Server
[22:03:09] Completed 19500000 out of 50000000 steps (39%).

Can anyone help? It cuts my PPD down to 4.5k each.

[slaveofconvention]

That random undefined error 99 for starting the core is what has killed all of my folding - none of my GPU's will do much more than 1k ppd when they were all doing between 5 and 12k ppd - unfortunately, I have no answers for you.

[DynamoNED]

Back before I was using FAH GPU Tracker V2, I ran into that exact same error using the GUI client from Stanford. At first, a complete restart of the Folding@Home client would fix the problem temporarily. After a while, though, it stopped reporting WU's completed, which is why I dumped it in favor of GPU Tracker.

Are you using FAH GPU Tracker? If so, then you can clear and reset your Folding@Home core simply closing the program and restarting it. If not, or if it is a continuous problem, though, you may want to try updating or re-downloading the client. If that doesn't fix it, you may want to start looking at driver issues. If you are using the newest nVidia drivers, they don't always play nice with Folding@Home clients; try rolling back your nVidia drivers back a version or two. Past that, I'd start looking for a hardware issue.

[mDust]

Previous termination of core was improper.
When you shut down the client, do you stop the cpu and gpus prior to closing the program or shutting down the OS? I know that they need to be stopped first or bad things can happen. It looks to me like the client was terminated in the middle of working at 38% and then it couldn't figure out how to resume properly.
Even if you're properly shutting it down, the driver could be failing and restarting which is practically the same thing as pulling the plug as far as the F@H client is concerned. Although I've never gotten folding errors, my ATI driver sometimes fails and recovers...or doesn't recover, and requires a hard reboot.:mad: It's pretty rare, but it does happen sometimes. This driver failure stops the card from doing anything and obviously doesn't politely warn the F@H client first.
You can find out if the driver is jacking up by doing the following (assuming you're running windows):
win+r>'eventvwr.msc'>enter
select 'windows logs' on the left>'applications'
Find the corresponding time in both logs and see if there is an error or warning in the windows log pertaining to your graphics driver.

[Fuganater]

Well the windows logs are clean. I am using GPU Tracker V2. I am actually using a 2v old driver I just noticed. I am downloading the newest to see how it goes.

[Twigsoffury]

Have you tried....Hitting it?

[Fuganater]

Haha no I havn't.

I installed the new 275 driver from NVIDIA and it seems to be a bit better. GPU1 still got that 99 error and droped to 8k PPD. Not as bad as 4k but still its not my normal 13-18k

[Yazymaria]

Well I am pretty impressed with the “new 275 driver from NVIDIA” as well, So I agree with “Fuganater” opinion.

[Fuganater]

So when it goes to crap I have to delete the WU, restart my PC and restart the client.... I wonder if it really is a GPU Tracker V2 issue.

[mDust]

So when it goes to crap I have to delete the WU, restart my PC and restart the client.... I wonder if it really is a GPU Tracker V2 issue.

I wouldn't think that restarting the computer would have any effect on the closed program unless it somehow leaves bad data in RAM indefinitely.

Is anyone else here folding with the same card/driver?