How to Build a Super Computer From Ten PCs

By Lester Godsey

i Thinkstock Images/Comstock/Getty Images

Super computers have come about due to demand for high-capacity computational power. Applications such as medical research and weather analysis are just two examples of real-world application of such technology. According to Top 500 Supercomputer Sites, more than 78 percent of the top 500 super computers in the world run Linux, a open source operating system. Because open source software allows access to the code behind the program, this solution is very popular with super computer projects.

Step 1

Set up the main computer, or node. If the 10 computers are not identical, determine the most powerful computer as this should be what controls the cluster. Install Linux on this computer and install the features you want. While there are a number of packages to choose from the Linux software, the only ones you are required to install to set up the cluster are NFS (network file system) and SSH (secure shell). Specifically, NFS makes it easier to share files between systems, affecting setup of the super computer. SSH is a secure, safe way to remotely connect from one computer to another, critical in the case of setting up a 10-computer super computer cluster.

Step 2

Install MPI (message passing interface) on your computer. This is used to allow the main computer control of the other computers in the super computer cluster. LAM is a popular choice and like many distributions of Linux, it is an open source option. Depending on your hardware, you might be able to find a pre-compiled version of the MPI software you have selected. If not, you will need to compile it yourself or get other hardware that is supported.

Step 3

Set up the member or slave computers. The only Linux software you should need is NFS and SSH. While you can install additional packages, chances are good that you will never need access to them from this member node as you will be working through the main computer once the super computer cluster is operational.

Step 4

Verify installation. Reboot all the machines and see if each one comes on. See if you can connect to each node from the master. If the hardware appears to be running as designed, invoke your MPI application, whether LAM or another messaging passing interface software.