If you want to get a head start on things we can go ahead and install cuda. This setup has been tested on 3 different linux distros and it works.
I try to keep things that depend on each other in the same location so I make a couple of folders just to keep it simple. You can just copy and paste the lines within the code tags and have it done.
Code:
mkdir ~/folding && mkdir -p ~/folding/cpu && mkdir -p ~/folding/gpu && cd ~/folding/gpu
You must use the 32 bit cuda toolkit regardless of whether you are on 32 or 64 bit. Download it and move it to the ~/folding/gpu folder you created.
Code:
http://www.nvidia.com/object/thankyou_linux.html?url=/compute/cuda/2.0-Beta2/linux/toolkit/NVIDIA_CUDA_Toolkit_2.0beta2_Ubuntu7.10_x86.run
Make it executable.
Code:
chmod +x NVIDIA_CUDA_Toolkit_2.0beta2_Ubuntu7.10_x86.run
Execute it with root privileges.
Code:
sudo ./NVIDIA_CUDA_Toolkit_2.0beta2_Ubuntu7.10_x86.run
Do not change the default install location, just hit enter when asked.
I'm going to have to do a little changing of the scripts to suit a Debian/Ubuntu based install.
In the meantime you can use the commands I found on standfords site..
http://fire-salamander.co.uk/uploads...eadless_nv.txt