"Getting to grips with..." supercomputing

The Formula 1 of the computing world

Supercomputing or high performance computing is the Formula 1 of the computing world. It lies in the specialist domain where cutting-edge computing equipment and associated software components – predominantly open source in nature – come together to push the boundaries of what can be achieved in terms of processing, networking, storage and visualisation. Access to supercomputing is helping local researchers address challenges in priority areas for South Africa, such as climatic change, energy security and communicable diseases.

Much like Formula 1 racing cars, which are designed to make it possible for drivers to compete at high speed, supercomputing allows users to step on the gas pedal of calculation speed.

Supercomputers help us solve problems today instead of tomorrow. Their ability to perform computations very rapidly and handle extremely large volumes of data effectively makes them invaluable for research projects where large volumes of data must be processed.

The speed of a supercomputer is generally measured in FLOPS (FLoating Point Operations Per Second), commonly used with an SI (International System of Units) or metric prefix such as tera-, combined into the shorthand TFLOPS (1012 FLOPS, pronounced teraflops), or peta-, combined into the shorthand PFLOPS (1015 FLOPS, pronounced petaflops). By comparison, the speed of a desktop is around 10 gigaflops.

The first supercomputers introduced in the 1960s were designed primarily by Seymour Cray at Control Data Corporation (CDC), and led the market into the 1970s. Today, some of the top supercomputers are typically one-of-a-kind custom designs produced by a range of companies.

Formula 1 drivers are trained to handle high-speed racing cars and get the best out of them. Similarly, supercomputers need specialists to coax the best results out of their considerable computing power. Dr Happy Sithole, director of the CHPC explains, “High performance computing is a powerful tool used by researchers in many different domains. It allows researchers to experiment by using simulations. It also speeds up the time required to process huge amounts of data.”

Given the current speed of progress, supercomputers are projected to reach 1 exaflops (1018) (one quintillion FLOPS) in 2019. Futurist Ray Kurzweil expects supercomputers capable of human brain neural simulations, for which, according to Kurzweil, 10 exaflops (1019) would be required in 2025.

Did you know? Some facts about supercomputers