The Ferranti Mark 1

Background

In 1948 the government's chief scientist initiated a government contract with Ferranti Ltd. to make a production version of the Manchester Mark 1 "to Professor Williams' specification". This was important because it established a link between the university and industry, and ensured that the power of the electronic computer would be available as quickly as possible around the U.K (and abroad).

The architecture was closely based on the Manchester Mark 1, which was effectively treated as the prototype for the commercial version. The main improvements (apart from improved engineering) were increased B-line, CRT, and magnetic drum stores, a much faster multiplier, and an increased range of instructions, for example many more B-line operations. The magnetic drum store and the multiplier were substantially redesigned.

The work on the many improvements on the Ferranti Mark 1 was shared between Ferranti personnel and University, with perhaps the biggest contribution coming from the two key members of the University staff who joined Ferranti in 1949, Geoff Tootill and especially Alec Robinson. Robinson redesigned his multiplier from essentially a serial multiplier to a much more parallel one, with an increase of speed of the order of 5 times on average. Coupled with the increased basic speed (from around 1.8 msec. per instruction to 1.2 msec.), and the more powerful B-line facilities, this roughly doubled the speed of the machine for scientific applications.

The Ferranti Mark 1 was the world's first commercially available general-purpose computer. The first machine off the production line was delivered to the University in February 1951. It was replaced in 1958.

After the second production Mark 1 was sold, further enhancements were made, and it was now called the Mark 1*. Ferranti publicly sold 9 production Mark 1/1*s between 1951 and 1957 (but one or two other machines might have gone to government agencies). Three machines were sold abroad (Canada, Holland and Italy).

A Ferranti sales brochure from 1952 can be found in the Mark 1 Literature. Here you can see a picture of a nicely packaged machine, a layman's guide to computers, and a list of useful things you can do with a computer, e.g., Determinants and Matrices, Ordinary Differential Equations, Partial Differential Equations, Function Tabulation, Logical problems (e.g. chess), Commercial Applications (e.g. wages), and Computer Diagnostics and Programming Aids. There is also a section giving a number of facts and figures.

Specification

The Ferranti Mark 1 specification was very similar to the final Manchester Mark 1 specification. A return was made to only one page per CRT to improve reliability. The specification is given below with the changes/improvements on the Ferranti Mark 1 highlighted.

• Store organised in 20-bit addressable "line"s, an instruction taking one line and a number two consecutive lines (see layout)
• Serial 40-bit arithmetic, with hardware add, subtract and multiply (with a double-length accumulator) and logical instructions
• 8 modifier registers (B-lines, for modifying addresses in instructions); simple B-line arithmetic and tests
• Single-address format order code with about 50 function codes
• 8 pages of random access main store (one CRT per 64 * 20-bit line page)
• 512 page capacity drum backing store, 2 pages per track, about 30 milliseconds revolution time
• Standard instruction time: 1.2 milliseconds, multiplication 2.16 milliseconds
• Peripheral Instructions : read and punch a line of 5-hole paper tape; transfer a given page (or track) on drum to/from a given Williams-Kilburn Tube "page" (or page pair) in store

The enhanced order code included a number of new "one-off" instructions (e.g. a random number generator, and an instruction counting the number of 1s in a word), a set of instructions to make the 8 B-lines more generally useful, and a few improvements/additions to the Arithmetic and Control instructions.

Input was by 5-hole paper tape, output was to a paper-tape punch or direct to a teleprinter for printing on-line. There were only 2 page display monitors, which could be switched to any of the RAM pages. (See pictures of the console and the input/output equipment on it.)

There was a single drum, substantially re-engineered by Ferranti and with a capacity 8 times that of the latest of the Manchester Mark 1 drums. When the first machine was delivered to the University only a quarter of the 16,000 word capacity was in commission, but this was increased as time progressed.

The Ferranti Mark 1*

Programming on the Mark 1

/E@A:SIU½DRJNFCKTZLWHYPQOBG"MXV£

This meant that integers could be expressed as 8 characters, instructions as 4 and machine addresses as 2 characters.

A description of programming on the Ferranti Mark 1 is given separately, and there is a version of the first chapter of the Turing-Brooker Programming Manual.

Here are three original Ferranti Mark 1 programmes, contributed by Huw Owen Pritchard, Distinguished Research Professor Emeritus, York University, Canada. [Added 27/07/10 by Toby Howard]

In 1954 a Mark 1 "Autocode" created by Tony Brooker was made available. It was an "easy to learn scientific programming language". Most programmers preferred it even though it slowed the machine down considerably.

Some idea of its impact can be gained from Tommy Thomas's piece on the significance of Brooker's autocodes in taking the early Manchester machines into the market (1954 - 1960).