With the advent of multi-core processors and the recent stagnation in processor clock frequency scaling, parallel processing is becoming more and more important in the evolution of computation in any class. Single-threaded applications will no longer be able to rely on processor clock speed and Instruction Level Parallelism to gain speed, leaving only parallel processing as a solution to this physical and technological bottleneck. Unfortunately, parallel processing has typically been reserved to an elite class of programmers and developers due to the complex nature of parallel environments when it comes to configuration and use of tools for programming, debugging and performance assessment.

Building machines for High Performance Computing (HPC) and Scientific computing, especially in a parallel environment, is a non-trivial task. It requires intricate operating system knowledge and highly specialized tools, most of which tend to be permanently masked in the Gentoo tree mostly due to their complexity and cross-dependencies.

The goal of this project is therefore to provide both users and administrators the means to quickly build a Gentoo based Beowulf clustering solution without having to hand pick and configure all the low-level, system packages. Also, a set of consistent tools will be required to seamlessly switch compilers and core libraries as it is often required in scientific environments. Many default configuration settings as well as automatic configuration scripts and wizards will have to be coded to guide the administrator and users into successfully using (and abusing) their HPC machines.