A Quick Start to using the CS Department's MST GPU Cluster

1. ssh into r01ckhw2.device.mst.edu using PuTTy or any ssh client. You should
get your credentials by email. If you didn't, then contact me at ckhw2@mst.edu
As soon as you login, you will be automatically redirected to one of the 3 nodes
gpu1, gpu2, or gpu3 to ease the usage of the login node gpu0

2. You should already have a directory called NVIDIA_CUDA_SDK in your home
directory. Run ‘make’ inside it to compile all the sample projects.
> cd NVIDIA_CUDA_SDK
> make
This might take a while. You can also make individual projects from inside the
~/NVIDIA_CUDA_SDK/projects directory.

3. The binaries get stored at ~/NVIDIA_CUDA_SDK/bin/linux/release. Try running
some of these as:
> ./deviceQuery
> ./matrixMul
etc.

4. To start writing your own program, you can use the template project at
~/NVIDIA_CUDA_SDK/projects/template. It has a ready-made Makefile.