Installing on Linux

Linux is the best choice for folding (besides a PS 3) as the client was natively coded for Linux, not to mention that it is the most lightweight of the three OS's. Linux only officially supports the console SMP client, however you can get the console GPU client to run as well, allowing you to crank out a large amount of points if you have either an Nvidia graphics card that supports the CUDA instructions respectively. To simply run the client on demand, first download the client from the Stanford Downloads page:

http://folding.stanford.edu/English/Download

Then, extract the package, and run the fah binary by typing "./<name of binary>". Then, simply fill in your user name, user key, and team number you got on the About Folding@home page on this site. You can then set the other options however you wish, although the default options are fine if you if you are not sure how else to set them. I recommend disabling CPU affinity lock, enabling "Big" work units, allocating all your available RAM, and setting the "-smp" parameter. Whenever you want to have the client running, simply execute this binary in a shell and leave the window open. For a more automatic system, follow the steps below to have the client run at boot for maximum production.

If you want the client to automatically start each time you boot your computer, you need to do a few more steps. First download and extract the client just as you did before. I recommend placing the resulting folder that contains your Folding@home binaries somewhere where it will not be disturbed or moved, such as your home directory. Next, go ahead and run the initial configuration of your client by running the folding at home client with the -configonly parameter by typing "./<name of executable> -configonly". You will then answer the questions as stated above, however once you are finished the client will exit. Next, download this script to your computer:

LINK TO SCRIPT

Copy this script to your folding at home directory, and make it executable by typing

chmod +x <path to script>

in a shell. Download these two files to your computer:

LINK

Copy the first file to /etc/init.d/ on your computer by typing

sudo cp <path to file> /etc/init.d/

Then copy the second file to your /etc/rc2.d/ folder by typing:

sudo cp <path to file> /etc/rc2.d/

This will run your client at boot, so that anytime your computer is on the Folding@home client will be running to ensure maximum production.

For extra points, you can run the GPU client as well if you have either an Nvidia card that supports the CUDA instructions, respectively. If you want to run this client, download the windows xp/2000 32-bit GPU client and extract it. You need to install wine, so if you do not have it installed already type "sudo apt-get install wine" in a shell of your choice and your distribution should install wine without issue. Once wine has been installed, you need to grab the Nvidia CUDA toolkit from Nvidia. Simply go to www.nvidia.com and search for the cuda toolkit; download the 32-bit cuda toolkit bundle. Install this bundle by typing

sudo sh <path to cuda bundle>

Once the bundle has installed, link it by typing:

sudo sh -c "echo '/usr/local/cuda/lib' > /etc/ld.so.conf.d/cuda.conf" sudo ldconfig

Create the wine directory by typing

wine notepad

and then exiting notepad once it has opened. Download the latest wine wrapper and link it by typing:

"wget http://www.gpu2.twomurs.com/wrapper2ndgen/2.1/cudart.dll.so -O ~/.wine/drive_c/windows/system32/cudart.dll" ln -s ~/.wine/drive_c/windows/system32/cudart.dll ~/.wine/drive_c/windows/system32/nvcuda.dll

and then type:

ldd ~/.wine/drive_c/windows/system32/cudart.dll

to ensure everything is linked correctly. As a precautionary measure, type

sudo ldconfig

again to ensure everything is linked correctly.Next, copy and paste this script and call it "cudainit.sh"

#!/bin/bash

PATH=$PATH:/usr/local/cuda/bin

modprobe nvidia

if [ "$?" -eq 0 ]; then

N=`expr $N3D + $NVGA - 1`
for i in `seq 0 $N`; do
mknod -m 666 /dev/nvidia$i c 195 $i;
done

mknod -m 666 /dev/nvidiactl c 195 255

else
exit 1
fi

Make it executable by typing

chmod u+x cudainit.sh

and then run this as root by typing

sudo ./cudainit.sh

Finally, cd to your windows GPU client directory, and start the client by typing:

nice wine Folding@home-Win32-GPU.exe -verbosity 9 -forcegpu nvidia_g80

The client should then start folding. Currently, we have not put together anything to start the client at boot, however the method for the SMP client can be easily modified to run the GPU client with minimal effort. If enough people request it, I will put something together. For more detailed information on these clients, please explore the links on the About Folding@home page on this site.