EVGA

Hot!Guide to Build/Install/Setup your own Multi-GPU Linux Folding Rig

Page: < 12 Showing page 2 of 2
Author
Wreck3r
New Member
  • Total Posts : 56
  • Reward points : 0
  • Joined: 2011/12/07 05:23:41
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/28 08:50:19 (permalink)
Cool thanks
#31
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/28 14:16:59 (permalink)
Great guide. As of nVidia 384 (at least) though can't add gpu slots (single GTX 1080 ti here). Tried other drivers too, same story. GPU folding is not supported via CUDA anymore ... next is trying openCL drivers ... maybe they'd work...
#32
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/28 14:53:58 (permalink)
you my friend forgot step #7 or are installing the nvidia driver that doesnt have CUDA.


#33
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/28 15:13:19 (permalink)
Chris21010
you my friend forgot step #7 or are installing the nvidia driver that doesnt have CUDA.




I'd wish. Even though it does let me see the gpu, it stays "ready" forevermore. It never says "running". Funny thing is that this whole setup worked on Windows, but I find Windows inefficient which is why I switched to Mint ... still unable to make it work (installed cuda in particular), even reinstalled the OS and followed the steps exactly ... I don't know what's going on. Client-type I use advanced as I did on Windows and packets "small". I may try diff values, but I can't see why them be the one to prevent this from working...
#34
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/28 17:16:02 (permalink)
you may need to manually set the ID's in the config.xml file like this:
 
<slot id='1' type='GPU'>
<gpu-index v='0'/>
<cuda-index v='0'/>
<opencl-index v='0'/> </slot>
<slot id='2' type='GPU'>
<gpu-index v='1'/>
<cuda-index v='1'/>
<opencl-index v='1'/>
</slot>
<slot id='3' type='GPU'>
<gpu-index v='2'/>
<cuda-index v='2'/>
<opencl-index v='2'/>
</slot>
<slot id='4' type='GPU'>
<gpu-index v='3'/>
<cuda-index v='3'/>
<opencl-index v='3'/>
</slot>
<slot id='5' type='GPU'>
<gpu-index v='4'/>
<cuda-index v='4'/>
<opencl-index v='4'/> </slot>



#35
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/29 00:42:19 (permalink)
Chris21010
you may need to manually set the ID's in the config.xml file like this:
 
<slot id='1' type='GPU'>
 
 
 
 
 
 
 
<gpu-index v='0'/>
<cuda-index v='0'/>
<opencl-index v='0'/> </slot>
<slot id='2' type='GPU'>
<gpu-index v='1'/>
<cuda-index v='1'/>
<opencl-index v='1'/>
</slot>
<slot id='3' type='GPU'>
<gpu-index v='2'/>
<cuda-index v='2'/>
<opencl-index v='2'/>
 
 
 
 
 
 
 
</slot>
<slot id='4' type='GPU'>
<gpu-index v='3'/>
<cuda-index v='3'/>
<opencl-index v='3'/>
 
 
 
 
 
 
 
</slot>
<slot id='5' type='GPU'>
<gpu-index v='4'/>
<cuda-index v='4'/>
<opencl-index v='4'/> </slot>
 
 
 
 
 
 
 



Predictably, didn't work, I only have one gpu. But yes, manually setting the index was one of the first things I tried ... it's still waiting on "FahCore Run".
Normally I would think it's hardware, but it works exceptionally well under Windows. Next is (again) trying another version of nvidia drivers and it that (predictably) won't work, it is time to get rid of Mint altogether and go to something that the drivers would hopefully have better chance to work well... I'm outta of my wits at this point.
#36
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/29 06:12:30 (permalink)
That makes no sense... you wouldnt still happen to have the log would you?
post edited by Chris21010 - 2017/08/29 06:23:09


#37
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/08/29 08:08:15 (permalink)
Chris21010
That makes no sense... you wouldnt still happen to have the log would you?


I'd try to produce one (for the time being I'm back to Windows to double check that folding works perfectly well there, and indeed it does)... in a few hours' time (possibly less) I'd be back with a log...
#38
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 03:36:11 (permalink)
Chris21010
That makes no sense... you wouldnt still happen to have the log would you?


Sorry, for taking up so much time. I needed to make sure that my system is as close to yours as possible. Also to make sure that it is functional under windows. So I grabbed a second Ti from a friend, ran it for a whole day under Windows (both ran fine) and then re-tried your guide. Even used the drivers that were current at the time of your write-up. Still same issue, my slots even failed at some point. I attach the log: https://ufile.io/v4bpa (slow download -> unzip)
 
Also I did not play with coolbits one bit, nor do I installed TeamViewer, so I followed your steps *exactly* up to step 8 (sans the teamviewer installation). I can even give you a teamviewer connection if everything else fails, I'm determined to make this work. From reading the log I can see that it does not detect the cuda cores, even though I install the nvidia-381 drivers, even as I manually assigned the cuda slots... So what's your take on this?
 
post edited by Stevethegreat - 2017/09/01 03:40:11
#39
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 06:08:45 (permalink)
try driver 381.22-0ubuntu0-~gpu16.04.2 its the one i am currently running. you may need to go to the software manager again and open the PPA to see the full list of installs.


#40
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 14:21:10 (permalink)
Chris21010
try driver 381.22-0ubuntu0-~gpu16.04.2 its the one i am currently running. you may need to go to the software manager again and open the PPA to see the full list of installs.


 Yes! That's exactly the version I'm running (nvidia-381 = driver 381.22) and installed exactly though the way you've said within your guide. I mean I followed your guide exactly, even used the name "mintuser" lest the username could somehow change anything (it wouldn't).
BTW is the log telling you anything additional? (other than that CUDA cores are not found). Funny things is that I'm even using the GPUs as in your guide and generally did everything apart from coolbits (which IMO should not change anything).
 
Slots keep failing on me though all the same ... probably I'd need to install ubuntu 16 for a change? Dunno, I'm at my wits' end. Also I noticed that your coolbits part is written for GTX 1070s , have you -maybe- did things slightly differently with your your GTX 1080 Ti rig and that's why I'm getting the showstoppers (i.e. the GTX 1080Tis need something additional not written on the guide?).
 
If you "refollow" your guide using Mint Linux and driver 381.22, are you going to get the same result as me? (on your Ti rig)
#41
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 15:32:17 (permalink)
at this point the last thing i can think of is try installing the CUDA toolkit. F@H log states no CUDA, which there should be. i honestly do not know enough about linux to really help diagnose the true issue or to even see if the CUDA libraries are installed or not. i know this is your issue but not how to correct it. the xorg.conf shouldn't have play here, i believe, and the default one created at install should be fine.


#42
Stevethegreat
New Member
  • Total Posts : 18
  • Reward points : 0
  • Joined: 2016/08/08 07:13:35
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 17:03:51 (permalink)
Chris21010
at this point the last thing i can think of is try installing the CUDA toolkit. F@H log states no CUDA, which there should be. i honestly do not know enough about linux to really help diagnose the true issue or to even see if the CUDA libraries are installed or not. i know this is your issue but not how to correct it. the xorg.conf shouldn't have play here, i believe, and the default one created at install should be fine.



As it turns out the issue was with nvidia-381 (of all things). So I "reverted" to nvidia-384 (latest version of nvidia drivers) and everything worked. So the showstopper was following your guide *exactly*. I don't know why or how but nvdia-381 simply does not work on my system. So yeah, I'm happilly folding at this point. It took 3-4 days to realize that there was definitely something wrong with the drivers themselves. I am still not sure what, what I do know is that whatever it was installing nvidia-384 on top of nvidia-381 solved it...
 
So I'm happy owner of a Mint-based folding rig. Thanks for the guide, here's hoping for less inconsistent drivers in the future. Let this also be a lesson for future would-be folder, if cuda cores are not recognized, simply install the latest nvidia drivers on top... It worked for me, it would probably work for you too...
#43
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/01 18:44:27 (permalink)
Well glad you got it working as that was really starting to bug me.


#44
Ranmacanada
iCX Member
  • Total Posts : 454
  • Reward points : 0
  • Joined: 2011/09/22 10:44:47
  • Status: offline
  • Ribbons : 1
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/14 12:38:56 (permalink)
So finally got this build working, and it's unbelievable.  I have my 1080FTW folding on an A6-5200.  When it was in windows, the 1080FTW was starving, and was only getting 360k PPD, when I installed Mint, the card is doing gangbusters better than it was doing when it was in my Xeon.  I needed to split my cards up since the current drivers for the 1080Ti suck so badly for the 1080's for folding.  So I have the 1080's in 2 different machines, 1 offsite as a space heater (in windows), 1 in the basement as a space heater and my 1080Ti in my main rig.  I am flabergasted that Linux has such a difference in needs of processing power for feeding the cards.  WOW.

 

Asus X79 Sabertooth
Xeon E5-2670 
Fold for the CURE!
EVGA 1080 FTW
MSI 1080 Gaming X
EVGA 1080Ti Hybrid
Register with my code SN278SHMDN :)
 
#45
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/14 14:36:37 (permalink)
Ya, windows is just bloatware for just a folding rig. And stock settings it practically rivals an OC on windows. All a win win for linux, the only loss is the harder environment to get things setup right.


#46
QuintLeo
iCX Member
  • Total Posts : 365
  • Reward points : 0
  • Joined: 2016/04/16 23:05:09
  • Status: offline
  • Ribbons : 2
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/16 03:06:43 (permalink)
But once you get a setup working, it becomes almost trivial to clone it - if you use the same MB and semi-close on the GPUs.
 
 I <3 dd
 
 
 
 

I am no longer the friendly FAH competition - thank Vorsholk and his ABUSE of his Curecoin developer position for that.
 I am still big on MOOing though!
 
#47
Hughesy26
New Member
  • Total Posts : 14
  • Reward points : 0
  • Joined: 2017/03/27 05:17:39
  • Location: Wales
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/20 07:11:13 (permalink)
Thanks for this great guide Chris...my 4x 1080ti rig has gone from 3.8-4m PPD to 4.5-4.7m. It's like you have given me a GTX 1080 (in PPD terms) that requires no electricity to run!!
 
Going to start looking at transferring some other rigs over to linux over the weekend. Wish I had been seen this months ago!
 


#48
Ranmacanada
iCX Member
  • Total Posts : 454
  • Reward points : 0
  • Joined: 2011/09/22 10:44:47
  • Status: offline
  • Ribbons : 1
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/09/23 15:07:18 (permalink)
I think I found something interesting for your tutorial Chris. 
https://github.com/RunGp/Nvidiux
 
Nvidiux is a graphical python tool to overclock or underclock your nvidia gpu

 

Asus X79 Sabertooth
Xeon E5-2670 
Fold for the CURE!
EVGA 1080 FTW
MSI 1080 Gaming X
EVGA 1080Ti Hybrid
Register with my code SN278SHMDN :)
 
#49
txgunner
New Member
  • Total Posts : 1
  • Reward points : 0
  • Joined: 2007/12/11 08:34:55
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/01 18:41:49 (permalink)
I've followed this guide verbatim with basically the same hardware setup.  I'm using driver version 381.22-0ubuntu0-~gpu16.04.2  I get as far as successfully installing the FaH Control, but the FaH clients failed to install.
 
I get this on the install of the clients:
 
Failed to install package 'fahclient_7_.4.4_amd64-1.deb'
Terminal:
 
(Reading database ... 202783 files and directories currently installed.)
Preparing to unpack .../fahclient_7.4.4_amd64-1.deb ...
usermod: no changes
Unpacking fahclient (7.4.4) over (7.4.4) ...
Setting up fahclient (7.4.4) ...
update-rc.d: error: no runlevel symlinks to modify, aborting!
dpkg: error processing package fahclient (--install):
 subprocess installed post-installation script returned error exit status 1
Processing triggers for gnome-menus (3.13.3-6ubuntu3.1) ...
Processing triggers for desktop-file-utils (0.22-1ubuntu5) ...
Processing triggers for mime-support (3.59ubuntu1) ...
Processing triggers for systemd (229-4ubuntu12) ...
Processing triggers for ureadahead (0.100.0-19) ...
ureadahead will be reprofiled on next reboot
 
Any idea on where I screwed up?  Any help would be most appreciated.
#50
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/01 19:04:42 (permalink)
you may need to update the OS before you can install. that or try downloading it again as i recently had errors on install till i downloaded it again.


#51
freEVfoldinGA
New Member
  • Total Posts : 73
  • Reward points : 0
  • Joined: 2012/01/01 05:21:10
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/05 20:13:08 (permalink)
First off thank you for the guide, it's very helpful.
 
I am running into a problem though. In step 10 I'm getting the error message "Cannot find the file /home/fold/.config/autostart/fan1.desktop" in xed, my username on the machine is fold.
 
Some more info is that I'm only using one GPU (980ti) and no team view or dummy plug, just directly attached monitor at this point, folding seems to be working though it doesn't autostart with the machine. I've updated linux and installed the nvidia drivers (387.12) correctly as far as I can tell.
 
Any help would be appreciated, I'm trying to use this machine as a space heater but I need to be able to quiet the fan down some more for it to fit that role.
 


#52
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/06 05:51:53 (permalink)
if it says it doesn't exist you'll need to create the directory first. as for F@H not starting at boot, that is something i haven't ran into as its setup to do that be the default install. try giving it a reinstall and see what happens.


#53
freEVfoldinGA
New Member
  • Total Posts : 73
  • Reward points : 0
  • Joined: 2012/01/01 05:21:10
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/07 08:04:30 (permalink)
Thanks for your help. I believe I got everything I need working. I created the directory and got GPU fan and overclock control working. Folding was actually autostarting, I must've paused it before rebooting the time I looked so it autostarted paused.
post edited by freEVfoldinGA - 2017/10/07 08:07:05


#54
Hughesy26
New Member
  • Total Posts : 14
  • Reward points : 0
  • Joined: 2017/03/27 05:17:39
  • Location: Wales
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/11 08:41:27 (permalink)
I've followed this guide with 2 rigs now, 1 with 4 1080tis and the other with just 1. All running great for a month or so.
 
I just brought 2 new 1080tis to put in the second rig. However, when I did so it seemed to mess up folding. I now can't get any of the 3 gpus in this system to fold! I have the following log:
 
*********** Folding@home Client *************************
15:29:58: Website: http://folding.stanford.edu/
15:29:58: Copyright: (c) 2009-2014 Stanford University
15:29:58: Author: Joseph Coffland < joseph@cauldrondevelopment.com>
15:29:58: Args: --child --lifeline 1882 /etc/fahclient/config.xml --run-as
15:29:58: fahclient --pid-file=/var/run/fahclient.pid --daemon
15:29:58: Config: /etc/fahclient/config.xml
15:29:58:******************************** Build ********************************
15:29:58: Version: 7.4.4
15:29:58: Date: Mar 4 2014
15:29:58: Time: 12:02:38
15:29:58: SVN Rev: 4130
15:29:58: Branch: fah/trunk/client
15:29:58: Compiler: GNU 4.4.7
15:29:58: Options: -std=gnu++98 -O3 -funroll-loops -mfpmath=sse -ffast-math
15:29:58: -fno-unsafe-math-optimizations -msse2
15:29:58: Platform: linux2 3.2.0-1-amd64
15:29:58: Bits: 64
15:29:58: Mode: Release
15:29:58:******************************* System ********************************
15:29:58: CPU: Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz
15:29:58: CPU ID: GenuineIntel Family 6 Model 79 Stepping 1
15:29:58: CPUs: 12
15:29:58: Memory: 15.51GiB
15:29:58:Free Memory: 14.98GiB
15:29:58: Threads: POSIX_THREADS
15:29:58: OS Version: 4.4
15:29:58:Has Battery: false
15:29:58: On Battery: false
15:29:58: UTC Offset: 1
15:29:58: PID: 1884
15:29:58: CWD: /var/lib/fahclient
15:29:58: OS: Linux 4.4.0-53-generic x86_64
15:29:58: OS Arch: AMD64
15:29:58: GPUs: 3
15:29:58: GPU 0: NVIDIA:7 GP102 [GeForce GTX 1080 Ti] 11380
15:29:58: GPU 1: NVIDIA:7 GP102 [GeForce GTX 1080 Ti] 11380
15:29:58: GPU 2: NVIDIA:7 GP102 [GeForce GTX 1080 Ti] 11380
15:29:58: CUDA: Not detected
15:29:58:***********************************************************************
15:29:58:<config>
15:29:58: <!-- Network -->
15:29:58: <proxy v=':8080'/>
15:29:58:
15:29:58: <!-- Slot Control -->
15:29:58: <power v='full'/>
15:29:58:
15:29:58: <!-- User Information -->
15:29:58: <passkey v='********************************'/>
15:29:58: <team v='224497'/>
15:29:58: <user v='Hughesy26_ALL_13HwtENuzNjjg2GoNiKwXSYt8Q1GqURCHT'/>
15:29:58:
15:29:58: <!-- Folding Slots -->
15:29:58: <slot id='0' type='GPU'>
15:29:58: <paused v='true'/>
15:29:58: </slot>
15:29:58: <slot id='1' type='GPU'>
15:29:58: <paused v='true'/>
15:29:58: </slot>
15:29:58: <slot id='2' type='GPU'>
15:29:58: <paused v='true'/>
15:29:58: </slot>
15:29:58:</config>
15:29:58:Switching to user fahclient
15:29:58:Trying to access database...
15:29:58:Successfully acquired database lock
15:29:58:Enabled folding slot 00: PAUSED gpu:0:GP102 [GeForce GTX 1080 Ti] 11380 (by user)
15:29:58:Enabled folding slot 01: PAUSED gpu:1:GP102 [GeForce GTX 1080 Ti] 11380 (by user)
15:29:58:Enabled folding slot 02: PAUSED gpu:2:GP102 [GeForce GTX 1080 Ti] 11380 (by user)
15:31:28:FS00:Unpaused
15:31:28:WU03:FS00:Starting
15:31:28:WU03:FS00:Running FahCore: /usr/bin/FAHCoreWrapper /var/lib/fahclient/cores/fahwebx.stanford.edu/cores/Linux/AMD64/NVIDIA/Fermi/Core_21.fah/FahCore_21 -dir 03 -suffix 01 -version 704 -lifeline 1884 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
15:31:28:WU03:FS00:Started FahCore on PID 2760
15:31:28:WU03:FS00:Core PID:2764
15:31:28:WU03:FS00:FahCore 0x21 started
15:31:29:WU03:FS00:0x21:*********************** Log Started 2017-10-11T15:31:28Z ***********************
15:31:29:WU03:FS00:0x21:Project: 9431 (Run 825, Clone 2, Gen 156)
15:31:29:WU03:FS00:0x21:Unit: 0x000000b6ab436c9d586fdd3a96f30ae5
15:31:29:WU03:FS00:0x21:CPU: 0x00000000000000000000000000000000
15:31:29:WU03:FS00:0x21:Machine: 0
15:31:29:WU03:FS00:0x21:Reading tar file core.xml
15:31:29:WU03:FS00:0x21:Reading tar file integrator.xml
15:31:29:WU03:FS00:0x21:Reading tar file state.xml
15:31:29:WU03:FS00:0x21:Reading tar file system.xml
15:31:29:WU03:FS00:0x21:Digital signatures verified
15:31:29:WU03:FS00:0x21:Folding@home GPU Core21 Folding@home Core
15:31:29:WU03:FS00:0x21:Version 0.0.18
15:31:29:WU03:FS00:0x21:ERROR:126: Bad platformId size.
15:31:29:WU03:FS00:0x21:Saving result file logfile_01.txt
15:31:29:WU03:FS00:0x21:Saving result file log.txt
15:31:29:WU03:FS00:0x21:Folding@home Core Shutdown: BAD_WORK_UNIT
15:31:29:WARNING:WU03:FS00:FahCore returned: BAD_WORK_UNIT (114 = 0x72)
15:31:29:WU03:FS00:Sending unit results: id:03 state:SEND error:FAULTY project:9431 run:825 clone:2 gen:156 core:0x21 unit:0x000000b6ab436c9d586fdd3a96f30ae5
15:31:29:WU03:FS00:Uploading 5.50KiB to 171.67.108.157
15:31:29:WU03:FS00:Connecting to 171.67.108.157:8080
15:31:30:WU03:FS00:Upload complete
15:31:30:WU00:FS00:Connecting to 171.67.108.45:80
15:31:30:WU03:FS00:Server responded WORK_ACK (400)
15:31:30:WU03:FS00:Cleaning up
15:31:30:WU00:FS00:Assigned to work server 140.163.4.233
15:31:30:WU00:FS00:Requesting new work unit for slot 00: READY gpu:0:GP102 [GeForce GTX 1080 Ti] 11380 from 140.163.4.233
15:31:30:WU00:FS00:Connecting to 140.163.4.233:8080
15:31:31:WU00:FS00:Downloading 25.55MiB
15:31:37:WU00:FS00:Download 8.32%
15:31:43:WU00:FS00:Download 16.39%
15:31:49:WU00:FS00:Download 24.71%
15:31:55:WU00:FS00:Download 33.76%
15:32:00:Removing old file 'configs/config-20171011-134113.xml'
15:32:00:Saving configuration to /etc/fahclient/config.xml
15:32:00:<config>
15:32:00: <!-- Network -->
15:32:00: <proxy v=':8080'/>
15:32:00:
15:32:00: <!-- Slot Control -->
15:32:00: <power v='full'/>
15:32:00:
15:32:00: <!-- User Information -->
15:32:00: <passkey v='********************************'/>
15:32:00: <team v='224497'/>
15:32:00: <user v='Hughesy26_ALL_13HwtENuzNjjg2GoNiKwXSYt8Q1GqURCHT'/>
15:32:00:
15:32:00: <!-- Folding Slots -->
15:32:00: <slot id='0' type='GPU'/>
15:32:00: <slot id='1' type='GPU'>
15:32:00: <paused v='true'/>
15:32:00: </slot>
15:32:00: <slot id='2' type='GPU'>
15:32:00: <paused v='true'/>
15:32:00: </slot>
15:32:00:</config>
15:32:01:WU00:FS00:Download 41.34%
15:32:07:WU00:FS00:Download 48.19%
15:32:13:WU00:FS00:Download 55.28%
15:32:19:WU00:FS00:Download 61.64%
15:32:25:WU00:FS00:Download 69.96%
15:32:31:WU00:FS00:Download 78.28%
15:32:37:WU00:FS00:Download 86.10%
15:32:43:WU00:FS00:Download 94.66%
15:32:46:WU00:FS00:Download complete
15:32:46:WU00:FS00:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:11409 run:191 clone:6 gen:52 core:0x21 unit:0x000000418ca304e95987376e00dc88bf
15:32:46:WU00:FS00:Starting
15:32:46:WU00:FS00:Running FahCore: /usr/bin/FAHCoreWrapper /var/lib/fahclient/cores/fahwebx.stanford.edu/cores/Linux/AMD64/NVIDIA/Fermi/Core_21.fah/FahCore_21 -dir 00 -suffix 01 -version 704 -lifeline 1884 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
15:32:46:WU00:FS00:Started FahCore on PID 2779
15:32:46:WU00:FS00:Core PID:2783
15:32:46:WU00:FS00:FahCore 0x21 started
15:32:47:WU00:FS00:0x21:*********************** Log Started 2017-10-11T15:32:46Z ***********************
15:32:47:WU00:FS00:0x21:Project: 11409 (Run 191, Clone 6, Gen 52)
15:32:47:WU00:FS00:0x21:Unit: 0x000000418ca304e95987376e00dc88bf
15:32:47:WU00:FS00:0x21:CPU: 0x00000000000000000000000000000000
15:32:47:WU00:FS00:0x21:Machine: 0
15:32:47:WU00:FS00:0x21:Reading tar file core.xml
15:32:47:WU00:FS00:0x21:Reading tar file integrator.xml
15:32:47:WU00:FS00:0x21:Reading tar file state.xml
15:32:47:WU00:FS00:0x21:Reading tar file system.xml
15:32:47:WU00:FS00:0x21:Digital signatures verified
15:32:47:WU00:FS00:0x21:Folding@home GPU Core21 Folding@home Core
15:32:47:WU00:FS00:0x21:Version 0.0.18
15:32:47:WU00:FS00:0x21:ERROR:126: Bad platformId size.
15:32:47:WU00:FS00:0x21:Saving result file logfile_01.txt
15:32:47:WU00:FS00:0x21:Saving result file log.txt
15:32:47:WU00:FS00:0x21:Folding@home Core Shutdown: BAD_WORK_UNIT
15:32:47:WARNING:WU00:FS00:FahCore returned: BAD_WORK_UNIT (114 = 0x72)
15:32:47:WU00:FS00:Sending unit results: id:00 state:SEND error:FAULTY project:11409 run:191 clone:6 gen:52 core:0x21 unit:0x000000418ca304e95987376e00dc88bf
15:32:47:WU00:FS00:Uploading 5.50KiB to 140.163.4.233
15:32:47:WU00:FS00:Connecting to 140.163.4.233:8080
15:32:48:WU00:FS00:Upload complete
15:32:48:WU02:FS00:Connecting to 171.67.108.45:80
15:32:48:WU00:FS00:Server responded WORK_ACK (400)
15:32:48:WU00:FS00:Cleaning up
15:32:48:WU02:FS00:Assigned to work server 140.163.4.233
15:32:48:WU02:FS00:Requesting new work unit for slot 00: READY gpu:0:GP102 [GeForce GTX 1080 Ti] 11380 from 140.163.4.233
 
I've tried upgrading the driver from 384 to 387 to no avail. I've also tried reverting back to 384, again no dice.
 
Any help would be much appreciated!


#55
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/11 09:03:04 (permalink)
bad work unit typically means one of two things. your OC is too high or the GPU is bad. if i where you i would test all three of those GPU's individually without and OC and make sure there is no issue with any specific GPU. if they all run fine then your board may be bad.


#56
Hughesy26
New Member
  • Total Posts : 14
  • Reward points : 0
  • Joined: 2017/03/27 05:17:39
  • Location: Wales
  • Status: offline
  • Ribbons : 0
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/11 22:50:26 (permalink)
Wow thanks for the quick reply Chris. Your last comment about a bad board got me thinking and I remembered this board has a supplemental sata power connector for if you add more than 2 cards. I connected this and hey presto it worked! 
 
Thanks again


#57
Chris21010
FTW Member
  • Total Posts : 1298
  • Reward points : 0
  • Joined: 2006/05/03 07:26:39
  • Status: online
  • Ribbons : 8
Re: Guide to Building/Install/Setup of your own Multi-GPU Linux Folding Rig 2017/10/12 07:14:02 (permalink)
Ah, yea those 1080ti's are power hungry.


#58
Page: < 12 Showing page 2 of 2
Jump to:
  • Back to Mobile