Architectural Trends in Scientific Computing

From vacuum tubes, to transistors, to integrated circuits, to microprocessors - the laboratory has been leading (and often paving) the way in designing, implementing, and exploiting the most innovative high performance computing architectures and technologies available. In the scientific computing sector, five generations of computing architectures have evolved since the days of the ENIAC. These architectures range from serial, vector, shared memory, distributed memory, to today's commodity processor based clusters.

The following diagram illustrates the architectural trends and system capabilities for the lineage of BRL / ARL scientific computing platforms - from the ENIAC to the world-class capabilities of today's DoD Supercomputing Resource Center at the Army Research Laboratory.

Chart describing the architectural trends and system capabilities of BRL/ARL Scientific Computing systems - Eniac to present.

Last Update / Reviewed: September 1, 2010