High Performance Computing Initiative To Enhance Engineering Education

Abstract

The power and utility of personal computers continues to grow exponentially through (1) advances in computing capabilities through newer microprocessors, (2) advances in microchip technologies, (3) electronic packaging, and (4) cost effective gigabyte-size hard-drive capacity. The engineering curriculum must not only incorporate aspects of these advances as subject matter, but must also leverage technological breakthroughs to keep programs competitive in terms of their infrastructure (i.e., delivery mechanisms, teaching tools, etc.). An aspect of these computing advances is computer modeling and simulation of engineering problems. Many engineering problems require significant computing power, and some complex problems require massive computing power. An example of a complex problem is a model that combines several aspects of a flight vehicle. Such a model might include fluid-solid interaction, heat transfer and dynamic loading of structures, all of which are coupled. Such models can easily consume massive computing resources, such as a supercomputer. To provide a conventional supercomputer on a dedicated basis to our faculty and upper level students is not feasible. It is feasible, however, to provide computing power adequate for teaching and student research in the form of clustered personal computers. Clusters can be acquired over time as individual computer purchases and configured by our own departmental personnel. Parallel computing software to exploit the clusters is available for computer operating systems like Unix, Windows NT or Linux. Clusters also have the advantage that they can be used as stand alone computers in a laboratory environment when they are not operating as a parallel computer.