EDVAC (Electronic Discrete Variable Automatic Computer) was one of the earliest electronic computers. Unlike its predecessor the ENIAC, it was binary rather than decimal, and was designed to be a stored-program computer.
ENIAC inventors John Mauchly and J. Presper Eckert proposed the EDVAC's construction in August 1944. A contract to build the new computer was signed in April 1946 with an initial budget of US$100,000. EDVAC was delivered to the Ballistic Research Laboratory in 1949. The Ballistic Research Laboratory became a part of the US Army Research Laboratory in 1952.
Functionally, EDVAC was a binary serial computer with automatic addition, subtraction, multiplication, programmed division and automatic checking with an ultrasonic serial memory capacity of 1,000 34-bit words. EDVAC's average addition time was 864 microseconds and its average multiplication time was 2,900 microseconds.
Project and plan
ENIAC inventors John Mauchly and J. Presper Eckert proposed EDVAC's construction in August 1944, and design work for EDVAC commenced before ENIAC was fully operational. The design would implement a number of important architectural and logical improvements conceived during the ENIAC's construction and would incorporate a high-speed serial-access memory. Like the ENIAC, the EDVAC was built for the U.S. Army's Ballistics Research Laboratory at the Aberdeen Proving Ground by the University of Pennsylvania's Moore School of Electrical Engineering. Eckert and Mauchly and the other ENIAC designers were joined by John von Neumann in a consulting role; von Neumann summarized and discussed logical design developments in the 1945 First Draft of a Report on the EDVAC.
A contract to build the new computer was signed in April 1946 with an initial budget of US$100,000. The contract named the device the Electronic Discrete Variable Automatic Calculator. The final cost of EDVAC, however, was similar to the ENIAC's, at just under $500,000.
The EDVAC was a binary serial computer with automatic addition, subtraction, multiplication, programmed division and automatic checking with an ultrasonic serial memory capacity of 1,000 44-bit words (later 1,024 words, thus giving a memory, in modern terms, of 5.5 kilobytes).
Physically, the computer comprised the following components:
- a magnetic tape reader-recorder (Wilkes 1956:36 describes this as a wire recorder.)
- a control unit with an oscilloscope
- a dispatcher unit to receive instructions from the control and memory and direct them to other units
- a computational unit to perform arithmetic operations on a pair of numbers at a time and send the result to memory after checking on a duplicate unit
- a timer
- a dual memory unit consisting of two sets of 64 mercury acoustic delay lines of eight words capacity on each line
- three temporary tanks each holding a single word
EDVAC's average addition time was 864 microseconds (about 1,160 operations per second) and its average multiplication time was 2,900 microseconds (about 340 operations per second). Time for an operation depended on memory access time, which varied depending on the memory address and the current point in the serial memory's recirculation cycle.
The computer had almost 6,000 vacuum tubes and 12,000 diodes, and consumed 56 kW of power. It covered 490 ft² (45.5 m²) of floor space and weighed 17,300 pounds (8.7 short tons; 7.8 t). The full complement of operating personnel was thirty people per eight-hour shift.
Impact on future computer design
John Von Neumann's famous EDVAC monograph, First Draft of a Report on the EDVAC, proposed the main enhancement to its design that embodied the principal "stored-program" concept that we now call the Von Neumann architecture. This was the storing of the program in the same memory as the data. The British computers EDSAC at Cambridge and the Manchester Baby were the first working computers that followed this design. And it has been followed by the great majority of all computers ever made since. Having the program and data in different memories is now called the Harvard architecture to distinguish it.
Installation and operation
EDVAC was delivered to the Ballistics Research Laboratory in 1949. After a number of problems had been discovered and solved, the computer began operation in 1951 although only on a limited basis.
In 1952 (April/May) it was running over 7 hours a day (period from 15 April to 31 May, used for 341 hours).
By 1957 EDVAC was running over 20 hours a day with error-free run time averaging 8 hours. EDVAC received a number of upgrades including punch-card I/O in 1954, extra memory in slower magnetic drum form in 1955, and a floating-point arithmetic unit in 1958.
- Wilkes, M. V. (1956). Automatic Digital Computers. New York: John Wiley & Sons. pp. 305 pages. QA76.W5 1956.
- "First Draft of a Report on the EDVAC" Archived 2004-04-23 at the Wayback Machine (PDF format) by John von Neumann, Contract No.W-670-ORD-4926, between the United States Army Ordnance Department and the University of Pennsylvania. Moore School of Electrical Engineering, University of Pennsylvania, June 30, 1945. The report is also available in Stern, Nancy (1981). From ENIAC to UNIVAC: An Appraisal of the Eckert–Mauchly Computers. Digital Press.
- Weik, Martin H. (December 1955). "EDVAC". ed-thelen.org. A Survey of Domestic Electronic Digital Computing Systems.
- Williams, Michael R. (1993). "The origins, uses, and fate of the edvac". IEEE Ann. Hist. Comput. 15 (1): 30, 32–33, 36–37. CiteSeerX 10.1.1.705.4726. doi:10.1109/85.194089.
- "The EDVAC". Digital Computer Newsletter. 3 (1): 2. April 1951.
- "6. Aberdeen Proving Ground Computers: The EDVAC". Digital Computer Newsletter. 4 (3): 4. July 1952.
- Moore School of Electrical Engineering (1949). A Functional Description of the EDVAC: A Report of Development Work under Contract W36-034-ORD-7593 with the Ordnance Department, Department of the Army. Philadelphia: University of Pennsylvania. A complete technical description of EDVAC's original structure and operation in 1949; includes an errata dated 1950. Fully viewable online.
- Oral history interview with J. Presper Eckert, Charles Babbage Institute, University of Minnesota.
- Oral history interview with Carl Chambers, Charles Babbage Institute, University of Minnesota.
- Oral history interview with Irven A. Travis, Charles Babbage Institute, University of Minnesota.
- Oral history interview with S. Reid Warren, Charles Babbage Institute, University of Minnesota.
- Oral history interview with Frances E. Holberton, Charles Babbage Institute, University of Minnesota.