High-performance computing: Past, present, and future

Abstract

Although the power of “yesterday’s“ supercomputers is now available on the desktop, our yearning for even more computational power to solve even larger problems continues to grow. This paper covers the highlights of some of the author's High Performance Computing (HPC) experiences dating back one of the first commercial supercomputers - Intel IPSC hypercube in the mid-80's - through today where her current HPC work focuses on cluster computing using MPI. The author will also give some of her prediction on where she thinks the HPC field is heading in the future. Her “Ant Theory” (even sugar ants display more complex planning and data processing than any humanbuilt systemof today) tells us we have a long way to go. Through newer technologies such as nanotubes and Bose-Einstein condensates, our future should even prove Moore's law wrong, and let the HPC field tackle even bigger and more complex problems.