The Computer as van Planned It
Original Publication Date: 1993-Apr-30
Included in the Prior Art Database: 2005-Oct-07
Software Patent Institute
M.D. GODFREY: AUTHOR [+3]
We describe the computer defined in von Neumann's unpublished paper ";First Draft of a Report on the EDVAC,"; Moore School of Electrical Engineering, University of Pennsylvania, June 30, 1945. We discuss motivation for the architecture and design, and contrast the machine with the EDVAC that was actually constructed.
THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.
Copyright ©; 1992 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Used with permission.
The Computer as van Planned It
We describe the computer defined in von Neumann's unpublished paper "First Draft of a Report on the EDVAC," Moore School of Electrical Engineering, University of Pennsylvania, June 30, 1945. We discuss motivation for the architecture and design, and contrast the machine with the EDVAC that was actually constructed.
John von Neumann made a key contribution to the understanding and development of computer architecture and design in his unpublished report titled "First Draft of a Report on the EDVAC." I However, in reading work that refers to this report and to the EDVAC computer (Eckert and Mauchly say the acronym stands for Electronic Discrete VAriable Computer-) which it described, some perplexing observations emerge:
-- The constructed EDVAC is usually described as being based on the von Neumann report.1
-- The von Neumann report is often described as the collective work of the Moore School group, unfairly given the sole authorship of von Neumann. (See, for example, Aspray and Burks' edition of the Papers of John con Neumann on Computers and Computer Theory, p. XV.3) This would suggest that many of the ideas in the report were shared by the Moore School design group and therefore would be expected to appear in the constructed machine.
-- The EDVAC has been referred to on numerous occasions, but these references do not agree about basic facts. For example, a key feature of any computer is the size of each word in the addressable memory. On this subject, Goldstine4 indicates 40 bits, others (for example, Burks5) say 32 bits. The only known publication giving the correct value (44 bits) is Knuth's articled The BRL Report7 (which is well known but was never published) also has the correct value.
Some of the evident confusion stems from the failure to distinguish between the "EDVAC" described by von Neumann in the report and the "EDVAC" constructed at the Moore School. While copies of the von Neumann report were informally circulated at the time it was written, the Moore School design documents were kept private and were in fact classified and marked "CONFIDENTIAL." (Eckert and Mauchly's "Automatic High- Speed Computing: A Progress Report on the EDVAC"2 was changed to "unclassified" in 1947.) Subsequently, the confusion has been aggravated by the fact that von Neumann's report has been reprinted only in incomplete or inaccurate forms.
The main purpose of this article is to present the architecture given in the von Neumann report in a form that is accessible to a wider audience and to translate into modern terminology the formal machine definition given in the von Neumann report. We also compare this definition with the definition of the constructed EDVAC system. In doing this, we hope to clarif...