• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

GPU Client linux guide

E

Enigma

[H]ard|Gawd
Joined
Oct 15, 2003
Messages
1,030
Looking for a GPU guide for linux.... specific to ubuntu 10.04 X64


I am looking at adding a GTX 480 to the garden can anyone help?

thanks in advance
 
kaltertod@Cyclops:~/GPUFAH$ wine Folding@home-Win32-GPU.exe -verbosity 9 -forcegpu nvidia_g80

Note: Please read the license agreement (Folding@home-Win32-GPU.exe -license). Further
use of this software requires that you have read and accepted this agreement.

[13:47:39] Gpu type=2 species=22.


--- Opening Log file [December 4 13:47:39 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.30r1

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\kaltertod\GPUFAH
Executable: Z:\home\kaltertod\GPUFAH\Folding@home-Win32-GPU.exe
Arguments: -verbosity 9 -forcegpu nvidia_g80

[13:47:39] - Ask before connecting: No
[13:47:39] - User name: enigma1983 (Team 33)
[13:47:39] - User ID: 1E9739562E989037
[13:47:39] - Machine ID: 2
[13:47:39]
[13:47:39] Gpu type=2 species=22.
[13:47:39] Loaded queue successfully.
[13:47:39]
[13:47:39] + Processing work unit
[13:47:39] Core required: FahCore_15.exe
[13:47:39] - Autosending finished units... [December 4 13:47:39 UTC]
[13:47:39] Trying to send all finished work units
[13:47:39] Core found.
[13:47:39] + No unsent completed units remaining.
[13:47:39] - Autosend completed
[13:47:39] Working on queue slot 01 [December 4 13:47:39 UTC]
[13:47:39] + Working ...
[13:47:39] - Calling '.\FahCore_15.exe -dir work/ -suffix 01 -nice 19 -checkpoint 15 -verbose -lifeline 8 -version 630'

Warning: Ignoring unknown arg
Warning: Ignoring unknown arg
[13:47:39]
[13:47:39] *------------------------------*
[13:47:39] Folding@Home GPU Core
[13:47:39] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[13:47:39]
[13:47:39] Build host: SimbiosNvdWin7
[13:47:39] Board Type: NVIDIA/CUDA
[13:47:39] Core : x=15
[13:47:39] Window's signal control handler registered.
[13:47:39] Preparing to commence simulation
[13:47:39] - Ensuring status. Please wait.
^Ckaltertod@Cyclops:~/GPUFAH$ wine Folding@home-Win32-GPU.exe -verbosity 9 -forcpu nvidia_fermi

Note: Please read the license agreement (Folding@home-Win32-GPU.exe -license). Further
use of this software requires that you have read and accepted this agreement.

[13:47:50] Gpu type=3 species=22.


--- Opening Log file [December 4 13:47:50 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.30r1

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\kaltertod\GPUFAH
Executable: Z:\home\kaltertod\GPUFAH\Folding@home-Win32-GPU.exe
Arguments: -verbosity 9 -forcegpu nvidia_fermi

[13:47:50] - Ask before connecting: No
[13:47:50] - User name: enigma1983 (Team 33)
[13:47:50] - User ID: 1E9739562E989037
[13:47:50] - Machine ID: 2
[13:47:50]
[13:47:50] Gpu type=3 species=22.
[13:47:50] Loaded queue successfully.
[13:47:50]
[13:47:50] + Processing work unit
[13:47:50] Core required: FahCore_15.exe
[13:47:50] Core found.
[13:47:50] Working on queue slot 01 [December 4 13:47:50 UTC]
[13:47:50] + Working ...
[13:47:50] - Calling '.\FahCore_15.exe -dir work/ -suffix 01 -nice 19 -checkpoint 15 -verbose -lifeline 34 -version 630'

[13:47:50] - Autosending finished units... [December 4 13:47:50 UTC]
[13:47:50] Trying to send all finished work units
[13:47:50] + No unsent completed units remaining.
[13:47:50] - Autosend completed
Getting Loaded...
Reading file work/wudata_01.tpr, VERSION 4.0.5 (single precision)
Warning: Ignoring unknown arg
Warning: Ignoring unknown arg
Note: tpx file_version 58, software version 59
Loaded with Money

[13:47:50]
[13:47:50] *------------------------------*
[13:47:50] Folding@Home GPU Core
[13:47:50] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[13:47:50]
[13:47:50] Build host: SimbiosNvdWin7
[13:47:50] Board Type: NVIDIA/CUDA
[13:47:50] Core : x=15
[13:47:50] Window's signal control handler registered.
[13:47:50] Preparing to commence simulation
[13:47:50] - Ensuring status. Please wait.
[13:47:50] Client config found, loading data.
Posq: cudaMalloc in CUDAStream::Allocate failed CUDA version is insufficient for CUDART version
^Ckaltertod@Cyclops:~/GPUFAH$ ^C
kaltertod@Cyclops:~/GPUFAH$ ^C
 
What driver version are you using, and what cuda toolkit version? The cuda toolkit and driver version are related - to use driver version X, you need toolkit version Y. I have no idea what X and Y are, though. You may just need either a newer toolkit or a newer video driver.
 
You need the 3.0 toolkit. From the guide Tobit linked:

6. Download the latest NVIDIA Linux 64-bit driver and the CUDA 3.0 32-bit toolkit:

wget http://us.download.nvidia.com/XFree86/Linux-x86_64/256.35/NVIDIA-Linux-x86_64-256.35.run
wget http://developer.download.nvidia.co.../cudatoolkit_3.0_linux_32_ubuntu9.04.runWatch the word wrap on the second command.

7. Before you install the NVIDIA driver, you need to install some Linux build utilities and header files:

sudo apt-get install build-essential linux-headers-`uname -r`
Those are backward apostrophes surrounding the uname command, which allows the output of uname to be incorporated into the filename of the headers file. Cut and paste from the above line if you have trouble downloading the proper headers.

8. Install the NVIDIA driver:

sudo sh NVIDIA-Linux-x86_64-256.35.run
Follow the prompts. Go ahead and install the OpenGL libraries too. Answer "Yes" when asked if you would like to run the nvidia-xconfig utility.

9. Install the NVIDIA 32-bit toolkit:

sudo sh cudatoolkit_3.0_linux_32_ubuntu9.04.run
Use the default path by hitting Enter when prompted.

10. Link the toolkit:

sudo sh -c "echo '/usr/local/cuda/lib' > /etc/ld.so.conf.d/cuda.conf"
sudo ldconfig
 
Note that if there is a newer toolkit than 3.0, you may need that with your driver version. Or, just use driver version 256.35 as the guide suggests.
 
same error as above with these directions, will try rolling back driver
 
I believe we figured it out..... stupid me forgot to delete the stanford supplied .dll's after creating the symlinks....... lol they were conflicting

Will Let you guys know how it goes after a little while
 
TPF on the 6811 seems a little high...... I adjusted my smp client on the box from using 6 cores to using just 5 it will slow the smp down a little bit but I am hoping to get more performance out of the GPU client.

I forgot how hot a gtx 480 gets when it is aircooled lol 87C Full load but should give the team a nice ppd boost.
 
Got my TPF down to about 8 mins. with the gtx 480 stock......

It appears that coolbits in linux does not work with 400 series cards yet as it does not unlock the clock frequencies in the control panel...... Anyone have a solution? nvclock does not work either........... no OC joy :(
 
Enigma said:
It appears that coolbits in linux does not work with 400 series cards yet as it does not unlock the clock frequencies in the control panel...... Anyone have a solution? nvclock does not work either........... no OC joy :(
Click to expand...
Yeah, you are gonna be SOL for now until coolbits is updated. When I used to fold on GPUs on headless Linux, I used to first find the most stable OC in Windows and then flash the OC settings directly to the card. I've never done it on a Fermi card however, only GTX280 and older.
 
thanks for the reply tobit...... I kinda figured that was the way it was going to go....

I can flash the bios I am just leary doing it to a $400+ vid card lol, But after all we are on the [H]ardforum not the oftforum
 
sorry to bring this back from the dead but how exactly do we delete (or know) the stanford supplied .dll's?

I am receiving an error i believe may be related to this

Code: 
wisp@ubuntu:~$ cd fahgpu3
wisp@ubuntu:~/fahgpu3$ ./auto-mator-fermi.sh

Note: Please read the license agreement (Folding@home-Win32-GPU.exe -license). Further 
use of this software requires that you have read and accepted this agreement.

[05:00:38] Gpu type=3 species=30.


--- Opening Log file [May 25 05:00:38 UTC] 


# Windows GPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.30r1

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\wisp\fahgpu3
Executable: C:\windows\system32\Folding@home-Win32-GPU.exe
Arguments: -forcegpu nvidia_fermi 

[05:00:38] - Ask before connecting: No
[05:00:38] - User name: WISP (Team 24)
[05:00:38] - User ID: 1BDE5EB0096BB93E
[05:00:38] - Machine ID: 2
[05:00:38] 
[05:00:38] Gpu type=3 species=30.
[05:00:38] Loaded queue successfully.
[05:00:38] 
[05:00:38] + Processing work unit
[05:00:38] Core required: FahCore_15.exe
[05:00:38] Core found.
[05:00:38] Working on queue slot 01 [May 25 05:00:38 UTC]
[05:00:38] + Working ...
Warning: Ignoring unknown arg 
Warning: Ignoring unknown arg 
[05:00:38] 
[05:00:38] *------------------------------*
[05:00:38] Folding@Home GPU Core
[05:00:38] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[05:00:38] 
[05:00:38] Build host: SimbiosNvdWin7
[05:00:38] Board Type: NVIDIA/CUDA
[05:00:38] Core      : x=15
[05:00:38]  Window's signal control handler registered.
[05:00:38] Preparing to commence simulation
[05:00:38] - Ensuring status. Please wait.
[05:00:48] - Looking at optimizations...
[05:00:48] - Working with standard loops on this execution.
[05:00:48] - Previous termination of core was improper.
[05:00:48] - Going to use standard loops.
[05:00:48] - Files status OK
[05:00:48] sizeof(CORE_PACKET_HDR) = 512 file=<>
[05:00:48] - Expanded 18421 -> 75731 (decompressed 411.1 percent)
[05:00:48] Called DecompressByteArray: compressed_data_size=18421 data_size=75731, decompressed_data_size=75731 diff=0
[05:00:48] - Digital signature verified
[05:00:48] 
[05:00:48] Project: 11214 (Run 1, Clone 64, Gen 38)
[05:00:48] 
[05:00:48] Entering M.D.
Getting Loaded...
Reading file work/wudata_01.tpr, VERSION 4.0.5 (single precision)
Note: tpx file_version 58, software version 59
Loaded with Money

Posq: cudaMalloc in CUDAStream::Allocate failed CUDA version is insufficient for CUDART version
[05:00:50] Tpr hash work/wudata_01.tpr:  130402956 322021594 2030305943 2344161705 935081316
[05:00:50] Working on 264 Fs_coil
[05:00:50] Client config found, loading data.
[05:00:53] CoreStatus = 63 (99)
[05:00:53] + Error starting Folding@home core.
[05:00:58] 
[05:00:58] + Processing work unit
[05:00:58] Core required: FahCore_15.exe
[05:00:58] Core found.
[05:00:58] Working on queue slot 01 [May 25 05:00:58 UTC]
[05:00:58] + Working ...
Warning: Ignoring unknown arg 
Warning: Ignoring unknown arg 
[05:00:58] 
[05:00:58] *------------------------------*
[05:00:58] Folding@Home GPU Core
[05:00:58] Version 2.15 (Tue Nov 16 09:05:18 PST 2010)
[05:00:58] 
[05:00:58] Build host: SimbiosNvdWin7
[05:00:58] Board Type: NVIDIA/CUDA
[05:00:58] Core      : x=15
[05:00:58]  Window's signal control handler registered.
[05:00:58] Preparing to commence simulation
[05:00:58] - Ensuring status. Please wait.
^CExiting, thank's for using auto-mator
wisp@ubuntu:~/fahgpu3$

Enigma said:
I believe we figured it out..... stupid me forgot to delete the stanford supplied .dll's after creating the symlinks....... lol they were conflicting

Will Let you guys know how it goes after a little while
Click to expand...
 
Enigma said:
same error as above with these directions, will try rolling back driver
Click to expand...

there are wrappers to help point stuff in the right direction...
I last gpu folded in linux with a 9800gx2 its been awhile....

edit... nvermind
 
You must log in or register to reply here.
Back
Top