Page 1 of 1
question about ppd
Posted: Thu Oct 03, 2013 2:13 pm
by Stile42
Hi, I am a new folder but fold with another long time folder. I have 2 machined each containing 2 gtx780 gpu's. One machine nets around 270k ppd but the other only about 53k ppd. I have configured them both for client-type beta and they have both been getting almost exclusively 8900 wu's. The machine getting 270k ppd has been folding since early August, then 2nd has folded only since mid September. Both machine OS's are windows 7 64-bit and both machines report near identical TPF of just over 3 minutes. Any ideas why Im getting such a huge ppd difference here? Each machine has folded well over 10 of the 8900 wu's so I don't think there can be any problem there. If my estimates are even close, I have lost a few million points and would sure like to rectify this. Thanks for any help!
Re: question about ppd
Posted: Thu Oct 03, 2013 5:05 pm
by Jesse_V
Sorry, we can't offer support for beta workunits outside of the beta team forum.
That being said, if I encountered this with regular production workunits, I would check to make sure that both machines use a passkey and that bonus points are coming in.
Re: question about ppd
Posted: Thu Oct 03, 2013 5:26 pm
by 7im
Hello Stile42, welcome to the forum.
Please post the *** System *** section of the fah log, and about 50 lines following it that shows the config and startup info. Thanks.
Re: question about ppd
Posted: Thu Oct 03, 2013 6:19 pm
by PantherX
Please note that you might want to change the beta client-type to advanced since there was an issue some time ago where a beta FahCore_17 was unstable thus all WUs completed by it were rejected by the Server. To fix this, you should Pause the GPU Slot(s), delete the FahCore_17 from your system, the default location would be %APPDATA%\FAHClient\cores\www.stanford.edu\~pande\Win32\AMD64\NVIDIA\Fermi\beta\Core_17.fah and then Unpause it so that it will download the correct version (make sure that your client-type is set to advanced).
Re: question about ppd
Posted: Thu Oct 03, 2013 7:24 pm
by Stile42
Ok, I switched to 'client-type advanced' so I can stay here now? Deleted fahcore_17 while paused and returned to folding - no effect on ppd - no bonus points. Will I notice a change when I get new wu's?
I hope this is what you wanted from the log file, I *'ed out username and team number - other than that, its verbatim from the file
Code: Select all
aving configuration to config.xml
19:50:40:<config>
19:50:40: <!-- Folding Core -->
19:50:40: <cpu-affinity v='true'/>
19:50:40:
19:50:40: <!-- Folding Slot Configuration -->
19:50:40: <power v='off'/>
19:50:40:
19:50:40: <!-- Network -->
19:50:40: <proxy v=':8080'/>
19:50:40:
19:50:40: <!-- Slot Control -->
19:50:40: <pause-on-battery v='false'/>
19:50:40:
19:50:40: <!-- User Information -->
19:50:40: <passkey v='********************************'/>
19:50:40: <team v='******'/>
19:50:40: <user v='********'/>
19:50:40:
19:50:40: <!-- Folding Slots -->
19:50:40: <slot id='0' type='GPU'>
19:50:40: <client-type v='advanced'/>
19:50:40: </slot>
19:50:40: <slot id='1' type='GPU'>
19:50:40: <client-type v='advanced'/>
19:50:40: </slot>
19:50:40: <slot id='2' type='CPU'>
19:50:40: <cpus v='-1'/>
19:50:40: <pause-on-start v='true'/>
19:50:40: </slot>
19:50:40:</config>
20:01:36:WU01:FS00:Downloading core from http://www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_17.fah
20:01:36:WU01:FS00:Connecting to www.stanford.edu:80
20:01:37:WU01:FS00:FahCore 17: Downloading 2.51MiB
20:01:43:WU01:FS00:FahCore 17: 77.16%
20:01:44:WU01:FS00:FahCore 17: Download complete
20:01:44:WU01:FS00:Valid core signature
20:01:44:WU01:FS00:Unpacked 8.48MiB to cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_17.fah/FahCore_17.exe
20:01:44:WU01:FS00:Starting
20:01:44:WU01:FS00:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/GMPC-2013/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_17.fah/FahCore_17.exe -dir 01 -suffix 01 -version 703 -lifeline 3576 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
20:01:44:WU01:FS00:Started FahCore on PID 2384
20:01:45:WU01:FS00:Core PID:4252
20:01:45:WU01:FS00:FahCore 0x17 started
20:01:45:WU02:FS01:Starting
20:01:45:WU02:FS01:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:/Users/GMPC-2013/AppData/Roaming/FAHClient/cores/www.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_17.fah/FahCore_17.exe -dir 02 -suffix 01 -version 703 -lifeline 3576 -checkpoint 15 -gpu 1 -gpu-vendor nvidia
20:01:45:WU02:FS01:Started FahCore on PID 1908
20:01:45:WU02:FS01:Core PID:3716
20:01:45:WU02:FS01:FahCore 0x17 started
20:01:51:WU02:FS01:0x17:*********************** Log Started 2013-10-02T20:01:50Z ***********************
20:01:51:WU02:FS01:0x17:Project: 8900 (Run 543, Clone 1, Gen 149)
20:01:51:WU02:FS01:0x17:Unit: 0x000000b4028c126651a69f781f6bee73
20:01:51:WU02:FS01:0x17:CPU: 0x00000000000000000000000000000000
20:01:51:WU02:FS01:0x17:Machine: 1
20:01:51:WU02:FS01:0x17:Digital signatures verified
20:01:51:WU02:FS01:0x17:Folding@home GPU core17
20:01:51:WU02:FS01:0x17:Version 0.0.49
20:01:51:WU02:FS01:0x17: Found a checkpoint file
20:01:51:WU01:FS00:0x17:*********************** Log Started 2013-10-02T20:01:50Z ***********************
20:01:51:WU01:FS00:0x17:Project: 8900 (Run 708, Clone 0, Gen 177)
20:01:51:WU01:FS00:0x17:Unit: 0x000000d2028c126651a6c41723fce6aa
20:01:51:WU01:FS00:0x17:CPU: 0x00000000000000000000000000000000
20:01:51:WU01:FS00:0x17:Machine: 0
20:01:51:WU01:FS00:0x17:Digital signatures verified
20:01:51:WU01:FS00:0x17:Folding@home GPU core17
20:01:51:WU01:FS00:0x17:Version 0.0.49
20:01:51:WU01:FS00:0x17: Found a checkpoint file
20:02:37:Removing old file 'configs/config-20130926-220913.xml'
20:02:37:Saving configuration to config.xml
20:02:37:<config>
20:02:37: <!-- Folding Core -->
20:02:37: <cpu-affinity v='true'/>
20:02:37:
20:02:37: <!-- Folding Slot Configuration -->
20:02:37: <power v='full'/>
20:02:37:
20:02:37: <!-- Network -->
20:02:37: <proxy v=':8080'/>
20:02:37:
20:02:37: <!-- Slot Control -->
20:02:37: <pause-on-battery v='false'/>
20:02:37:
20:02:37: <!-- User Information -->
20:02:37: <passkey v='********************************'/>
20:02:37: <team v='******'/>
20:02:37: <user v='********'/>
20:02:37:
20:02:37: <!-- Folding Slots -->
20:02:37: <slot id='0' type='GPU'>
20:02:37: <client-type v='advanced'/>
20:02:37: </slot>
20:02:37: <slot id='1' type='GPU'>
20:02:37: <client-type v='advanced'/>
20:02:37: </slot>
20:02:37: <slot id='2' type='CPU'>
20:02:37: <cpus v='-1'/>
20:02:37: <pause-on-start v='true'/>
20:02:37: </slot>
20:02:37:</config>
20:04:38:WU02:FS01:0x17:Completed 500000 out of 2500000 steps (20%)
20:04:38:WU02:FS01:0x17:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
20:04:39:WU01:FS00:0x17:Completed 1150000 out of 2500000 steps (46%)
20:04:39:WU01:FS00:0x17:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
20:08:01:WU01:FS00:0x17:Completed 1175000 out of 2500000 steps (47%)
20:08:02:WU02:FS01:0x17:Completed 525000 out of 2500000 steps (21%)
Re: question about ppd
Posted: Thu Oct 03, 2013 7:52 pm
by bollix47
The TPFs look right for a 780 running a P8900. If you're going by the estimated PPD in FAHControl give it a few frames to adjust. If you're using HFM.net switch the setting for the option "Calculate PPD based on" to All Frames.
Also, your log shows you're still using the beta core which is the result of the current work unit being download with the beta option. That should rectify itself once you get your next work unit.
Remember that each core 17 slot requires a CPU core for itself. Assuming you have 4 or more cores adjust your number of CPUs for your CPU slot from -1 to the total number of available cores minus 2.
Re: question about ppd
Posted: Thu Oct 03, 2013 9:05 pm
by PantherX
Regarding your bonus points, since the WU was already downloaded with the correctly configured username and passkey, you should be able to get the bonus points unless you have fallen below the 80% return rate. If you want to check your passkey percentage, please PM one of the Administrators/Moderators with your username and passkey.
Re: question about ppd
Posted: Thu Oct 03, 2013 9:51 pm
by EXT64
Could you double check your Windows and BIOS clocks and make sure both are correct?
Re: question about ppd
Posted: Thu Oct 03, 2013 11:24 pm
by Stile42
WOOHOO!!!! Things looking much better now, thank you all so much for the help. Seems that keeping the BIOS and Windows clocks sync'ed is important after all
Re: question about ppd
Posted: Fri Oct 04, 2013 12:46 am
by EXT64
Great! Good news is that glitch should only mess up PPD predictions - the server should have awarded you the correct points.