[SPG_Active_Members] Comments on the invention of the Stored Program Computer

[Randall Neff]

Mike Williams gave a lecture to the docents on Feb 7, 2004 on ENIAC.
He described the 'missing link' between the original design of ENIAC and
the stored program computer design, EDVAC.

There is a reprint of paper by H. H. Goldstine and Adele Goldstine
from 1946 in IEEE Annals of the History of Computing, Vol 18, No 1, 1966.
[Available online for money]
See also:
http://en.wikipedia.org/wiki/ENIAC
and its links.

Originally, ENIAC was programmed by plugging cables and setting knobs and
switches.  There were three function tables which were arrays of 1200
ten position
switches that stored the numeric constants of the problem.
In this design, it was possible to have parallel operations executing.
The first part of the report is at:
http://ftp.arl.mil/~mike/comphist/46eniac-report/index.html

In the Spring of 1947, J. von Neumann suggested that it would be
possible to run the ENIAC in a way very different from the way
contemplated when it was designed.  First, the machine was
plugged into a standard dataflow configuration.
Additional logic was added that converted a pair of ten digit
function table switches into control signals for the machine.
So each of the 100 positions was an order (instruction).  Each function
table could have 600 orders, all three would be 1800.
Originally, there were 60 different instructions, later increased to
92.  Now, ENIAC could be programmed in an hour  or two, rather than
hours or days.
The report is at:
http://ftp.arl.mil/~mike/comphist/48eniac-coding/index.html

[Note that the CDC 6600 used toggle switches for the boot
program]

So programs were now lists of numbers stored in read-only
switches.  ENIAC read numeric data from punched cards.
If the program switches were replaced with read/write
elements (in the ENIAC, this would be at least ten dual
triad vacuum tubes per switch), then programs could be
read from punched cards as well.

This led to von Neumann's draft report on EDVAC:
http://www.virtualtravelog.net/entries/2003-08-TheFirstDraft.pdf
See also:
http://en.wikipedia.org/wiki/First_Draft_of_a_Report_on_the_EDVAC





Randall.