Supercomputer
The introduction of the microprocessor integrated circuit (IC) in 1971 gave birth to two new categories of computer, the microcomputer and the supercomputer. The microcomputer developed first, drawing on the sudden progress in miniaturization represented by the microprocessor. Shortly thereafter, as further scientific breakthroughs doubled the microprocessor's speed and complexity each year, the term "supercomputer" came to be used to designate whatever computer currently contained the largest number of ICs, had the highest speed, and (usually) sold for the highest price.
The Cray 1 Computer, designed by Seymour Cray, founder of Cray Research of Chippewa Falls, Wisconsin, is generally regarded as the first supercomputer. Containing two hundred thousand ICs, it operated at one hundred fifty megaflops, which means one hundred fifty million floating point calculations per second. A floating point calculation is the addition, subtraction, multiplication or division of two very large or small numbers of the type described in scientific notation, such as 1.345 x 1011. Such calculations require much more work than fixed-point or integer operations.
Supercomputers are used for the most calculation-intensive tasks such as oil and mineral prospecting, analysis of subatomic activity, studying the earth's changing ozone layer, and many computer design and simulation projects, including design of aircraft, microprocessors, and ships; and simulation of weather systems, the birth of the universe, nuclear reactions, DNA, and many other tasks.
Cray, who had previously worked for Univac and later helped found Control Data Corporation, made his goal the production of the world's most powerful computers. Until the late 1980s, Cray fulfilled that goal nearly all of the time, but with increasing competition in the U.S. and Japan. Cray and others achieved very powerful computing in several ways, some of which include:
1. Adding more circuits to single ICs--As scientists found more ways to create smaller and smaller transistors, capacitors, resistors and diodes within semiconductors, the complexity and speed of each single IC quickly increased.
2. Adding parallel processing--Scientists also realized that the fastest, most powerful computing device, the human brain, achieved its power not from ultrafast processors, but from having many processors operating simultaneously, or in parallel. This has led most supercomputer firms to design computers having two or more microprocessors. For instance, the Cray-2 uses four microprocessors and has a peak speed of about 2.2 gigaflops (2.2 billion floating point calculations per second).
3. Conversion to massively parallel processing--Based on the success of their initial efforts in parallel processing, most supercomputer firms, including Cray, Control Data, Ncube, Inc., NEC, Fujitsu, Hitachi, Thinking Machines, Intel, IBM and others, are designing supercomputers with thousands or even millions of microprocessor ICs operating simultaneously. Most have set a goal of producing a computer by the year 2000 capable of operating at one teraflop (one trillion floating point operations per second).
4. Use of gallium arsenide instead of silicon chips--Cray's new company, Cray Computer, Inc., of Colorado Springs, is attempting to build its Cray 3 Supercomputer using gallium arsenide chips, which operate at lower voltages, higher speeds, and produce less heat. The Cray 3 is designed to have a peak speed between twenty and thirty gigaflops.
In 1996, Cray merged with workstation giant Silicon Graphics, Inc. to jointly produce supercomputers.
This is the complete article, containing 525 words
(approx. 2 pages at 300 words per page).
