The ENIAC was built to undertake arithmetic rather than the logical calculations for which the COLOSSUS was designed.  ENIAC employed conditional branching, greatly adding to its functionality and allowing more complex programs to be run.


The basic work for which ENIAC was designed was the solution of partial differential equations, which occur in many areas of applied science, and many of which could not be solved prior to the advent of computers. In its first three months of operation it worked for some weeks on a preliminary design of the H-Bomb.


Like the COLOSSUS, ENIAC employed parallel processing. However, it was capable of much more complex and varied calculation than the COLOSSUS. The method of programming the ENIAC - which involved extensive cross-wiring between plug boards - was such that it might take a team of programmers a day or so to set up the program, that might then take only a few minutes to run. It was very reliable. Later, in 1947/48, the programmers found that the programs could be kept in the machine’s store originally designed only for numerical data, and it was used in this mode in March/April 1948.


This essentially made ENIAC the first stored-program electronic computer. However, the store was read-only (thus not allowing the ENIAC to achieve its full potential), an important distinction to bear in mind when we look at the later Manchester ‘Baby’. The performance was slower using a stored-program, but the increased functionality of programs more than made up for this fact.

THE FIRST NUMBER CRUNCHER - 2

John Presper Eckert  (first from left), Herman Goldstine  (fourth from left), John Mauchly  (fifth from left), Paul Gillon (first from right)