Multiple computers connected together via a high speed network and configured to process a single program are known as 'clusters'. Using commodity computers in this configuration was popularized in1994 by Donald Becker at NASA's Goddard Space Flight Center. Dubbed Beowulf this computing configuration has proven very successful at solving extremely large problems at affordable prices. It is especially popular within the University research, high energy physics and earth sciences communities.

Instructions and data are passed between each computer [node] in the cluster via Message Passing Interface (MPI) libraries. MPI is a library specification for message-passing, proposed as a standard by a broadly based committee of vendors, The primary MPI information repository is at Argonne National Labs.

The powerful yet inexpensive commodity computers [especially Intel based] and free Linux operating system have accelerated development of cluster configurations to the point where specialized Linux distributions designed specifically for clusters are now being developed.

Absoft offers a version of Pro Fortran designed especially for cluster users. It includes floating Fortran F95/F77 licenses plus Beowulf Software Toolkit which includes a variety of pre-built cluster software.

The links below provide additional information and resources for those who wish to learn more about clusters.

Cluster Resources

• MPI Standard
• Beowulf's Home page
• Beowulf Underground
• Scyld - 2nd generation software for cluster users
• WMPI - MPI for Windows users
• Appleseed - MPI software for Mac OS9 and OSX users
• MPI under OS X - UCSD Supercomputer Center
• LAM/MPI - Local Area Multi-computer
• LAM/MPI - Mac OS X Help guide
• mpich - Mac OS X Help Guide
• mpich - Home page
• MPICH/MPICH2 Build instructions for your OSX, 32 & 64 bit Linux systems.
• Pooch MPI Mac OS X
  
• MPI Software Technology
• Distributed Debugging Tool
• Black Lab Cluster Management - PowerPC
• Monthly Bulletin Subsciption