Supercomputers are the most powerful computing machines on the planet. They use thousands of processor chips, cost millions of dollars and fill small buildings. Minicomputers are much smaller midsized machines that fall between mainframes and personal computers in processing speed, cost and power. Various types of minicomputers are manufactured in large numbers with standard components. In contrast, supercomputers are rare and they are often custom-built to specifications.
Minicomputers manage networks of computers and other devices. In the 1960s, Massachusetts Institute of Technology used the first minicomputer to control a network of telephone lines. Other early minicomputers functioned as brains for networks of dumb terminals, with all software run by the terminal operators located on the minicomputers. Today, personal computers capable of running their own software have replaced dumb terminals. High end PCs sometimes function as minicomputers, providing services across a network.
Scientists use supercomputers to tackle questions about the origins of the universe, the behavior of weather and other problems requiring massive amounts of mathematical calculation. The National Oceanic and Atmospheric Administration’s Oak Ridge supercomputer is helping researchers understand climate change. Deep Blue, an IBM machine designed to play chess like a grand master, beat world chess champion Gary Kasparov in 1997. Supercomputers are being used to figure out mathematical algorithms to help reduce nuclear reactor waste.
Size and Speed
Minicomputers are stand-alone machines similar to PCs in size. The Roadrunner supercomputer takes up about 5,000 square feet at the Los Alamos National Laboratory in New Mexico. Minicomputers measure their processor speeds in gigahertz (GHz), or billions of cycles per second. Top performing minicomputer processors run at 3 GHz and higher. Roadrunner has a top speed of just over a quadrillion (1,000,000,000,000,000) calculations per second.