Is my GPU working correctly

[silverpulser]

HP p6-2005ukm, 4Gb ram, CPU Processor Intel(R) Pentium(R) CPU G620 @ 2.60GHz, 2 Core(s), 2 Logical Processor(s), GPU NVIDIA GeForce GT 520 (48 CUDA cores) with 1Gb RAM on board, driver version 10.18.13.5906 dated 24/11/2015. OS Windows 10 Pro, 64 bit.

The assigned work unit to the GPU seems to be running slow - less than 10% per day. The first three or four days were only run at 12 hours per day but for the past week it has been running 24 hours per day. Looking at the Web Control page it says that the unit ETA is 5.44 days but the other dialogue says that it has 20 hours 20 mins to complete the work unit. Does this mean I have left my computer running for 24/7 for no good reason and that the work unit will lapse in 20 hours? Is there something wrong with my setup that is not making full use of the GPU?

http://1drv.ms/1Pfp4cL

[JonasTheMovie]

Are the slots manually configured or set to -1? Is the slider set to anything else than Full?
In "auto" config the medium setting makes FAH wait for he GPU to be fully idle to use it.

[bruce]

You probably know that if you want a computer to run 24/7 you have to disable some of the power-saving features such as hibernate/sleep.

If you post your log (see my sig) we can give you a lot better impression of what might be going on.

[Joe_H]

A GT 520 is a fairly low end GPU and may not be able to complete many of the current projects using Core_17 or later. Especially without running 24 hours a day. You do need to let the WU complete a few percent before the estimates will be accurate.

Your GT 520 should be able to complete the older Core_15 projects within their deadlines. If it can be confirmed that this GPU is not fast enough for later cores, and you want to try those projects, directions for setting up the GPU slot are available.

[silverpulser]

Are the slots manually configured or set to -1

Yes, the slots are set to -1 and the slider is set to Full

[silverpulser]

You probably know that if you want a computer to run 24/7 you have to disable some of the power-saving features such as hibernate/sleep.

Sleep mode is set to Never

If you post your log (see my sig) we can give you a lot better impression of what might be going on.

Here is the log for my GPU slot 1

Code: Select all

*********************** Log Started 2016-01-04T10:12:14Z ***********************
10:53:36:FS01:Unpaused
10:53:36:WU02:FS01:Starting
10:53:36:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
10:53:36:WU02:FS01:Started FahCore on PID 7752
10:53:36:WU02:FS01:Core PID:7720
10:53:36:WU02:FS01:FahCore 0x18 started
10:53:37:WU02:FS01:0x18:*********************** Log Started 2016-01-04T10:53:36Z ***********************
10:53:37:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
10:53:37:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
10:53:37:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
10:53:37:WU02:FS01:0x18:Machine: 1
10:53:37:WU02:FS01:0x18:Digital signatures verified
10:53:37:WU02:FS01:0x18:Folding@home GPU core18
10:53:37:WU02:FS01:0x18:Version 0.0.4
10:53:37:WU02:FS01:0x18:  Found a checkpoint file
10:53:53:WU02:FS01:0x18:Completed 7850000 out of 16000000 steps (49%)
10:53:53:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
13:47:07:WU02:FS01:0x18:Completed 8000000 out of 16000000 steps (50%)
******************************* Date: 2016-01-04 *******************************
16:52:11:WU02:FS01:0x18:Completed 8160000 out of 16000000 steps (51%)
19:56:42:WU02:FS01:0x18:Completed 8320000 out of 16000000 steps (52%)
******************************* Date: 2016-01-04 *******************************
22:49:39:FS01:Paused
22:49:39:FS01:Shutting core down
22:49:39:WU02:FS01:0x18:WARNING:Console control signal 1 on PID 7720
22:49:39:WU02:FS01:0x18:Exiting, please wait. . .
22:49:39:WU02:FS01:0x18:Lost lifeline PID 7752, exiting
22:49:39:WU02:FS01:0x18:ERROR:103: Lost client lifeline
22:49:39:WU02:FS01:0x18:Folding@home Core Shutdown: CLIENT_DIED
22:49:40:WU02:FS01:FahCore returned: INTERRUPTED (102 = 0x66)
22:50:51:FS01:Unpaused
22:50:51:WU02:FS01:Starting
22:50:51:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
22:50:52:WU02:FS01:Started FahCore on PID 8068
22:50:52:WU02:FS01:Core PID:1196
22:50:52:WU02:FS01:FahCore 0x18 started
22:50:53:WU02:FS01:0x18:*********************** Log Started 2016-01-04T22:50:52Z ***********************
22:50:53:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
22:50:53:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
22:50:53:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
22:50:53:WU02:FS01:0x18:Machine: 1
22:50:53:WU02:FS01:0x18:Digital signatures verified
22:50:53:WU02:FS01:0x18:Folding@home GPU core18
22:50:53:WU02:FS01:0x18:Version 0.0.4
22:50:53:WU02:FS01:0x18:  Found a checkpoint file
22:51:09:WU02:FS01:0x18:Completed 8450000 out of 16000000 steps (52%)
22:51:09:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
23:25:59:WU02:FS01:0x18:Completed 8480000 out of 16000000 steps (53%)
02:30:29:WU02:FS01:0x18:Completed 8640000 out of 16000000 steps (54%)
******************************* Date: 2016-01-05 *******************************
05:35:01:WU02:FS01:0x18:Completed 8800000 out of 16000000 steps (55%)
08:39:42:WU02:FS01:0x18:Completed 8960000 out of 16000000 steps (56%)
******************************* Date: 2016-01-05 *******************************
11:44:45:WU02:FS01:0x18:Completed 9120000 out of 16000000 steps (57%)
13:07:54:FS01:Paused
13:07:55:FS01:Shutting core down
13:07:55:WU02:FS01:0x18:WARNING:Console control signal 1 on PID 1196
13:07:55:WU02:FS01:0x18:Exiting, please wait. . .
13:07:56:WU02:FS01:0x18:Lost lifeline PID 8068, exiting
13:07:56:WU02:FS01:0x18:ERROR:103: Lost client lifeline
13:07:56:WU02:FS01:0x18:Folding@home Core Shutdown: CLIENT_DIED
13:07:56:WU02:FS01:FahCore returned: INTERRUPTED (102 = 0x66)
15:24:14:FS01:Unpaused
15:24:14:WU02:FS01:Starting
15:24:14:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
15:24:15:WU02:FS01:Started FahCore on PID 904
15:24:16:WU02:FS01:Core PID:7788
15:24:16:WU02:FS01:FahCore 0x18 started
15:24:16:WU02:FS01:0x18:*********************** Log Started 2016-01-05T15:24:16Z ***********************
15:24:16:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
15:24:16:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
15:24:16:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
15:24:16:WU02:FS01:0x18:Machine: 1
15:24:16:WU02:FS01:0x18:Digital signatures verified
15:24:16:WU02:FS01:0x18:Folding@home GPU core18
15:24:16:WU02:FS01:0x18:Version 0.0.4
15:24:16:WU02:FS01:0x18:  Found a checkpoint file
15:24:35:WU02:FS01:0x18:Completed 9150000 out of 16000000 steps (57%)
15:24:35:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
******************************* Date: 2016-01-05 *******************************
17:54:28:WU02:FS01:0x18:Completed 9280000 out of 16000000 steps (58%)
20:58:42:WU02:FS01:0x18:Completed 9440000 out of 16000000 steps (59%)
******************************* Date: 2016-01-05 *******************************
00:02:55:WU02:FS01:0x18:Completed 9600000 out of 16000000 steps (60%)
03:07:13:WU02:FS01:0x18:Completed 9760000 out of 16000000 steps (61%)
******************************* Date: 2016-01-06 *******************************
06:11:26:WU02:FS01:0x18:Completed 9920000 out of 16000000 steps (62%)
09:15:51:WU02:FS01:0x18:Completed 10080000 out of 16000000 steps (63%)
******************************* Date: 2016-01-06 *******************************
11:39:29:FS01:Paused
11:39:29:FS01:Shutting core down
11:39:29:WU02:FS01:0x18:WARNING:Console control signal 1 on PID 7788
11:39:29:WU02:FS01:0x18:Exiting, please wait. . .
11:39:29:WU02:FS01:0x18:Lost lifeline PID 904, exiting
11:39:29:WU02:FS01:0x18:ERROR:103: Lost client lifeline
11:39:29:WU02:FS01:0x18:Folding@home Core Shutdown: CLIENT_DIED
11:39:30:WU02:FS01:FahCore returned: INTERRUPTED (102 = 0x66)
11:45:03:FS01:Unpaused
11:45:03:WU02:FS01:Starting
11:45:03:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
11:45:03:WU02:FS01:Started FahCore on PID 7416
11:45:03:WU02:FS01:Core PID:2176
11:45:03:WU02:FS01:FahCore 0x18 started
11:45:04:WU02:FS01:0x18:*********************** Log Started 2016-01-06T11:45:04Z ***********************
11:45:04:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
11:45:04:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
11:45:04:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
11:45:04:WU02:FS01:0x18:Machine: 1
11:45:04:WU02:FS01:0x18:Digital signatures verified
11:45:04:WU02:FS01:0x18:Folding@home GPU core18
11:45:04:WU02:FS01:0x18:Version 0.0.4
11:45:04:WU02:FS01:0x18:  Found a checkpoint file
11:45:20:WU02:FS01:0x18:Completed 10200000 out of 16000000 steps (63%)
11:45:20:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
12:31:40:WU02:FS01:0x18:Completed 10240000 out of 16000000 steps (64%)
15:36:09:WU02:FS01:0x18:Completed 10400000 out of 16000000 steps (65%)
******************************* Date: 2016-01-06 *******************************
18:40:36:WU02:FS01:0x18:Completed 10560000 out of 16000000 steps (66%)
19:47:36:FS01:Paused
19:47:36:FS01:Shutting core down
19:47:36:WU02:FS01:0x18:WARNING:Console control signal 1 on PID 2176
19:47:36:WU02:FS01:0x18:Exiting, please wait. . .
19:47:37:WU02:FS01:0x18:Lost lifeline PID 7416, exiting
19:47:37:WU02:FS01:0x18:ERROR:103: Lost client lifeline
19:47:37:WU02:FS01:0x18:Folding@home Core Shutdown: CLIENT_DIED
19:47:38:WU02:FS01:FahCore returned: INTERRUPTED (102 = 0x66)
20:01:46:FS01:Unpaused
20:01:46:WU02:FS01:Starting
20:01:46:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
20:01:46:WU02:FS01:Started FahCore on PID 6812
20:01:47:WU02:FS01:Core PID:2176
20:01:47:WU02:FS01:FahCore 0x18 started
20:01:48:WU02:FS01:0x18:*********************** Log Started 2016-01-06T20:01:47Z ***********************
20:01:48:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
20:01:48:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
20:01:48:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
20:01:48:WU02:FS01:0x18:Machine: 1
20:01:48:WU02:FS01:0x18:Digital signatures verified
20:01:48:WU02:FS01:0x18:Folding@home GPU core18
20:01:48:WU02:FS01:0x18:Version 0.0.4
20:01:48:WU02:FS01:0x18:  Found a checkpoint file
20:02:04:WU02:FS01:0x18:Completed 10600000 out of 16000000 steps (66%)
20:02:04:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
22:20:41:WU02:FS01:0x18:Completed 10720000 out of 16000000 steps (67%)
******************************* Date: 2016-01-06 *******************************
01:24:56:WU02:FS01:0x18:Completed 10880000 out of 16000000 steps (68%)
04:29:12:WU02:FS01:0x18:Completed 11040000 out of 16000000 steps (69%)
******************************* Date: 2016-01-07 *******************************
07:33:28:WU02:FS01:0x18:Completed 11200000 out of 16000000 steps (70%)
08:36:05:FS01:Paused
08:36:05:FS01:Shutting core down
08:36:06:WU02:FS01:0x18:WARNING:Console control signal 1 on PID 2176
08:36:06:WU02:FS01:0x18:Exiting, please wait. . .
08:36:07:WU02:FS01:0x18:Lost lifeline PID 6812, exiting
08:36:07:WU02:FS01:0x18:ERROR:103: Lost client lifeline
08:36:07:WU02:FS01:0x18:Folding@home Core Shutdown: CLIENT_DIED
08:36:07:WU02:FS01:FahCore returned: INTERRUPTED (102 = 0x66)
08:37:39:FS01:Unpaused
08:37:41:WU02:FS01:Starting
08:37:41:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/User/AppData/Roaming/FAHClient/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/Core_18.fah/FahCore_18.exe -dir 02 -suffix 01 -version 704 -lifeline 2364 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
08:37:41:WU02:FS01:Started FahCore on PID 8176
08:37:41:WU02:FS01:Core PID:220
08:37:41:WU02:FS01:FahCore 0x18 started
08:37:42:WU02:FS01:0x18:*********************** Log Started 2016-01-07T08:37:42Z ***********************
08:37:42:WU02:FS01:0x18:Project: 9430 (Run 131, Clone 1, Gen 75)
08:37:42:WU02:FS01:0x18:Unit: 0x00000062ab40413855474df7ea894c96
08:37:42:WU02:FS01:0x18:CPU: 0x00000000000000000000000000000000
08:37:42:WU02:FS01:0x18:Machine: 1
08:37:42:WU02:FS01:0x18:Digital signatures verified
08:37:42:WU02:FS01:0x18:Folding@home GPU core18
08:37:42:WU02:FS01:0x18:Version 0.0.4
08:37:42:WU02:FS01:0x18:  Found a checkpoint file
08:37:58:WU02:FS01:0x18:Completed 11250000 out of 16000000 steps (70%)
08:37:58:WU02:FS01:0x18:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900

[7im]

I have 96 CUDA cores on my GT 430 compared to the 48 in the 520. My card gets better production running the older core_15 work units. Unfortunately, those will eventually run out as the project has moved on to newer versions.

Add the client-type setting of beta
Add the max-packet-size setting of small

How to add slot options is shown in the Configuration Guide on FAH web site.

[silverpulser]

I have 96 CUDA cores on my GT 430 compared to the 48 in the 520. My card gets better production running the older core_15 work units. Unfortunately, those will eventually run out as the project has moved on to newer versions.

Add the client-type setting of beta
Add the max-packet-size setting of small

How to add slot options is shown in the Configuration Guide on FAH web site.

So really what you are suggesting is that a newer/faster GPU is the answer in the longer term but in the meantime stick to core_15 work units. To be honest, I am surprised that the auto configuration setup doesn't do that, rather than send work units that are not likely to be completed in time. Would I be better off disabling the GPU and enable CPU-2?

[davidcoton]

So really what you are suggesting is that a newer/faster GPU is the answer in the longer term but in the meantime stick to core_15 work units. To be honest, I am surprised that the auto configuration setup doesn't do that, rather than send work units that are not likely to be completed in time. Would I be better off disabling the GPU and enable CPU-2?

No, even a weak GPU will beat the CPU hands down. As for auto-config coping with this -- Cores come and go, but the client goes on for ever. It can't possibly "know" about all the special cases, not because there are many (I think only this one, in fact) but because the situation arose after the code was written.

[7im]

The setup probably assumes folding will take place 24/7 or close to it, and sends work based on that assumption. The trick to fold core_15 is a deviation that was created purposefully to allow lower end and part time GPU folders to contribute to FAH.

And if memory serves, core_15 uses very little CPU, so set CPU:2 as well.

[silverpulser]

[quote="silverpulser"......

No, even a weak GPU will beat the CPU hands down. As for auto-config coping with this -- Cores come and go, but the client goes on for ever. It can't possibly "know" about all the special cases, not because there are many (I think only this one, in fact) but because the situation arose after the code was written.

In that case something is not quite right. The GPU project was, as predicted, dumped by FAH and a new and smaller WU assigned. The new WU (Core_18 and 3566 points) has been running for about 18 hours now and is showing as 9% completed. The PPD is only showing as 486 whereas my CPU-1 is showing as 1092 PPD.

I have now amended my configuration to the GPU slot as suggested to run core_15 WU's in future.

[davidcoton]

Did you set the options 7im gave? Core18 still has large WUs, you need the options to give you Core15 units that will complete on time.
It is possible that your GPU is the exception to what I wrote, but as 7im says, if you can get Core15 WUs on the GPU then you can run the CPU as well. The CPU (running both cores) will interfere with any core from 17 onwards.

[silverpulser]

Did you set the options 7im gave? Core18 still has large WUs, you need the options to give you Core15 units that will complete on time.
It is possible that your GPU is the exception to what I wrote, but as 7im says, if you can get Core15 WUs on the GPU then you can run the CPU as well. The CPU (running both cores) will interfere with any core from 17 onwards.

Yes I did. I expect the next assigned WU to be a core_15.

Incidentally, large difference in points between the WU that failed to complete (over 21000) and the WU just assigned (3566).


Is my current NVidia driver 359.06 known to be stable?

[7im]

We don't really track the quality of the almost daily released versions of drivers from NV. It's easier to track the rare unstable ones, and that version doesn't seem to be a problem.

[silverpulser]

Are there any special settings I should set in my NVidia control panel? At the moment everything is set at default.