Characteristics of first-generation computers

· Use of vacuum tubes in electronic circuits and mercury delay lines for memory'

· Magnetic drum as primary internal storage medium

· Limited main storage capacity (1000-4000 bytes)

· Low-level symbolic language programming

· Heat and maintenance problems

· Applications: scientific computations, payroll processing, re­cord keeping

· Cycle time: milliseconds

· Cost: $5 per floating-point operation

· Processing speed: 2000 instructions per second


 

TEXT 1 FIRST GENERATION COMPUTERS. HISTORY OF

DEVELOPMENT (PART 1)

1943-1947

First-generation computers were extremely large and had poor re­liability. They used vacuum tubes to control internal operations, gen­erated considerable heat, and required a lot of floor space. Although first-generation computers were much faster than earlier mechanical or electromechanical devices, they were very slow compared to to­day's computers, and their internal storage capacity was limited.

Punched cards were used to enter data into the computer. The holes were punched according to a coding scheme (much like Hol­lerith's cards), and a special-purpose machine (card reader) was used to translate them into machine languagefor the computer. The ma­chine language information was often stored on magnetic drums,cyl­inders coated with magnetizable material, rotating at high speeds. Read/write headssuspended just above the rotating surface of the drum either wrote on the drum by magnetizing small spots or read from it by interpreting the already magnetized spots. Numbers were manipulated by the computer according to the instructions, or pro­gram, given to it. The results of these operations were punched on blank cards, which could then be read by humans.

Only binary-coded machine language was used in early com­puters. With hardware costs dominating the developments of first- generation computers, the use of system software to relieve the user of low-level programming was just beginning.

1943.During 1940 and 1941 , Atanasoff and Berry met with John W. Mauchly and showed him their work. Mauchly, working at the Moore School of Electrical Engineering of the University of Pennsylvania, then began formulating his own ideas on how a general-purpose computermight be built. Mauchly’s ideas came to the attention of J. Presper Eckert, Jr. , a graduate engineering student at the Moore School, and the team of Mauchly and Eckert w'as formed. In 1943 they designed their first elec­tronic computer, ENIAC(Electronic Number Integrator and Calculator). It used 140 kilowatts of electricity and contained about 18,000 vacuum tubes, 70,000 resistors, and 10,000 capacitors, occupied a spmany components linked by close to a million hand-soldered connections. The I/O system consisted of modified IBM card readers and punches. ENIAC had a limited storage capacity of only 20 ten-digit numbers (it took 12 vacuum tubes to store one decimal digit), used a 100-kilohertz clock, and could perform 5000 additions or 300 multiplications per second. By today's standards, ENIAC was very slow; however, when delivered in 1946, it represented a major advance in computational power. It was in­strumental, for better or worse, in bringing the wwld into the atomic age.

1945.Dr. John von Neumann recommended in a research report that the binary number system, employing only the digits 0 and 1, be applied in computer design. He also proposed that instructions to control the com­puter, as well as data, should be stored within the computer. The EDSAC(Electronic Delay Storage Automatic Calculator), built at Cambridge Uni­versity in 1949, was the first computer to incorporate these ideas. It was not faster than ENIAC, but it did use the binary number system, and its in­structions were stored internally. These instructions were called a program; hence the name stored-program.

1947. The SSEC(Selective Sequence Electronic Calculator) was installed at IBM New York World Headquarters in 1947 and was used through 1952. At about the same time, IBM built a machine that could multiply six-digit numbers by counting electronic pulses. This ma­chine, which was simply a tabulating machine connected to some vac­uum tubes, was known as the IBM 603 electronic multiplier.