Difference between revisions of "Manchester Mark I"

From Computer History Wiki
Jump to: navigation, search
(Reasonably complete)
 
(More on name; slight re-factor to improve flow)
Line 1: Line 1:
The '''Manchester Mark I''' (its formal name at the time; it now often given as the '''Mark 1''' - it was also historically called the '''MADM''', or '''Manchester Automatic Digital Machine''', but this name is not still used) was a very early computer, built at the University of Manchester.
+
The '''Manchester Mark I''' (apparently its formal name at the time, although this is uncertain, as no ''images'' of any contemporaneous document by the builders giving the name they used seems to be readily available - it is now often given as the '''Mark 1'''; it was also historically called the '''MADM''', '''Manchester Automatic Digital Machine''', and '''MUC''', '''Manchester University Computer''', but these names are not used now) was a very early [[computer]], built at the University of Manchester.
  
It was constructed out of [[vacuum tube]]s; its [[arithmetic logic unit|ALU]] was [[serial]], used [[two's complement]], and had [[hardware]] support for multiply operations. It used a pair of [[Williams tube]]s for its [[main memory]]; each held two 'pages', each containing 32 40-[[bit]] [[word]]s, per tube. In addition to the Williams tube main memory, it also had a [[head]]-per-[[track]] [[drum]] (since providing a large amount of Williams tube main memory would have been difficult), with a revolution time of 30 milliseconds; it held 32 pages, with 2 pages per track.  
+
It was constructed out of [[vacuum tube]]s, and used a pair of [[Williams tube]]s for its [[main memory]]; each held two 'pages', each containing 32 40-[[bit]] [[word]]s per tube. In addition to the Williams tube main memory, it also had a [[head]]-per-[[track]] [[drum]] (since providing a large amount of Williams tube main memory would have been difficult), with a revolution time of 30 milliseconds; it eventually held 32 pages, with 2 pages per track.  
  
It was re-built, starting late in 1948, from the [[Manchester Baby]] (built to test the first Williams tubes). It went through two main phases itself, the Intermediary Version (operating by April, 1949), and the Final Specification (operating by October, 1949). The primary improvement between the two was that transfers to and from the drum could initially only be achieved under manual control; also added was the ability to do [[input/output|I/O]] via a [[paper tape]] reader/punch. (Since work on the machine was constant, various specification numbers - e.g. the size of the drum - varied over time.)
+
It was initially re-built, starting late in 1948, from the [[Manchester Baby]] (which was built to test the first Williams tubes). It went through two main phases itself, the Intermediary Version (operating by April, 1949), and the Final Specification (operating by October, 1949). The primary improvement between the two was that transfers to and from the drum could initially only be done under manual control; also added was the ability to do [[input/output|I/O]] via a [[paper tape]] reader/punch. (Since work on the machine was constant, various specification numbers - e.g. the size of the drum - varied over time.)
  
A production version was produced by [[Ferranti]], the [[Ferranti Mark I]]. The Manchester Mark I was shut down in the summer of 1950, to be replaced by a Ferranti Mark I in February, 1951.
+
A production version was produced by [[Ferranti]], the [[Ferranti Mark 1]]. The Manchester Mark I was shut down in the summer of 1950, to be replaced by a Ferranti Mark 1 in February, 1951.
  
 
==Architectural details==
 
==Architectural details==
  
The Mark I's biggest architectural improvement over the Baby was the addition of the first-ever [[index register]]s, called 'B-lines', two ('B0' and 'B1') were held in an additional Williams tube, the 'B-tube'. B0 was generally loaded with '0', to allow un-indexed memory references. Other Williams tubes held:
+
Its [[arithmetic logic unit|ALU]] was [[serial]], used [[two's complement]], and had [[hardware]] support for multiply operations.
  
* the [[accumulator]] ('A'), two 40-bit 'lines', in one (also 'A');
+
The Mark I's biggest architectural improvement over the Baby was the addition of the first-ever [[index register]]s, called 'B-lines'; two ('B0' and 'B1') were held in an additional Williams tube, the 'B-tube'. B0 was generally loaded with '0', to allow un-indexed memory references. Other Williams tubes held other [[register]]s:
 +
 
 +
* the [[accumulator]] (called 'A'), double-width, in two 40-bit 'lines', in one (also called the 'A');
 
* the [[Program Counter|program counter]] ('CI') and the [[Instruction Register|instruction register]] ('PI') in another ('C');
 
* the [[Program Counter|program counter]] ('CI') and the [[Instruction Register|instruction register]] ('PI') in another ('C');
 
* the multiplier ('R') and multiplicand ('D') [[operand]]s for multiplication, in the last ('M').
 
* the multiplier ('R') and multiplicand ('D') [[operand]]s for multiplication, in the last ('M').
  
 
The Mark I had 20-bit [[instruction]]s, stored two per word; it initially had 26 [[operation code]]s, later 30 (when the I/O operations were added). Instructions contained a 5-bit op-code, a 1-bit B-line selection, and a single memory [[address]].
 
The Mark I had 20-bit [[instruction]]s, stored two per word; it initially had 26 [[operation code]]s, later 30 (when the I/O operations were added). Instructions contained a 5-bit op-code, a 1-bit B-line selection, and a single memory [[address]].
 +
 +
''Note: There are inconsistencies between sources as to some of the details; take everything here with a grain of salt!''
  
 
{{semi-stub}}
 
{{semi-stub}}
Line 22: Line 26:
  
 
* Simon H. Lavington, ''A History of Manchester Computers'', National Computer Centre, Manchester, 1976
 
* Simon H. Lavington, ''A History of Manchester Computers'', National Computer Centre, Manchester, 1976
* Simon H. Lavington, ''Early British Computers'', Manchester University, Manchester, 1980
+
* Simon H. Lavington, ''[https://ed-thelen.org/comp-hist/EarlyBritish.html Early British Computers]'', Manchester University, Manchester, 1980
 
* Raúl Rojas, Ulf Hashagen (editors), ''The First Computers — History and Architectures'', MIT Press, Cambridge, 2000 - the Mark I is covered in Part IV-2, pp. 369-377
 
* Raúl Rojas, Ulf Hashagen (editors), ''The First Computers — History and Architectures'', MIT Press, Cambridge, 2000 - the Mark I is covered in Part IV-2, pp. 369-377
 
* B. V. Bowden (editor), ''Faster Than Thought: A Symposium on Digital Computing Machines'', Pitman Publishing, New York, 1964 - the Mark I is covered in Chapter 6, pp. 117-124; written by Tom Kilburn
 
* B. V. Bowden (editor), ''Faster Than Thought: A Symposium on Digital Computing Machines'', Pitman Publishing, New York, 1964 - the Mark I is covered in Chapter 6, pp. 117-124; written by Tom Kilburn

Revision as of 00:07, 14 March 2024

The Manchester Mark I (apparently its formal name at the time, although this is uncertain, as no images of any contemporaneous document by the builders giving the name they used seems to be readily available - it is now often given as the Mark 1; it was also historically called the MADM, Manchester Automatic Digital Machine, and MUC, Manchester University Computer, but these names are not used now) was a very early computer, built at the University of Manchester.

It was constructed out of vacuum tubes, and used a pair of Williams tubes for its main memory; each held two 'pages', each containing 32 40-bit words per tube. In addition to the Williams tube main memory, it also had a head-per-track drum (since providing a large amount of Williams tube main memory would have been difficult), with a revolution time of 30 milliseconds; it eventually held 32 pages, with 2 pages per track.

It was initially re-built, starting late in 1948, from the Manchester Baby (which was built to test the first Williams tubes). It went through two main phases itself, the Intermediary Version (operating by April, 1949), and the Final Specification (operating by October, 1949). The primary improvement between the two was that transfers to and from the drum could initially only be done under manual control; also added was the ability to do I/O via a paper tape reader/punch. (Since work on the machine was constant, various specification numbers - e.g. the size of the drum - varied over time.)

A production version was produced by Ferranti, the Ferranti Mark 1. The Manchester Mark I was shut down in the summer of 1950, to be replaced by a Ferranti Mark 1 in February, 1951.

Architectural details

Its ALU was serial, used two's complement, and had hardware support for multiply operations.

The Mark I's biggest architectural improvement over the Baby was the addition of the first-ever index registers, called 'B-lines'; two ('B0' and 'B1') were held in an additional Williams tube, the 'B-tube'. B0 was generally loaded with '0', to allow un-indexed memory references. Other Williams tubes held other registers:

  • the accumulator (called 'A'), double-width, in two 40-bit 'lines', in one (also called the 'A');
  • the program counter ('CI') and the instruction register ('PI') in another ('C');
  • the multiplier ('R') and multiplicand ('D') operands for multiplication, in the last ('M').

The Mark I had 20-bit instructions, stored two per word; it initially had 26 operation codes, later 30 (when the I/O operations were added). Instructions contained a 5-bit op-code, a 1-bit B-line selection, and a single memory address.

Note: There are inconsistencies between sources as to some of the details; take everything here with a grain of salt!

Further reading

  • Simon H. Lavington, A History of Manchester Computers, National Computer Centre, Manchester, 1976
  • Simon H. Lavington, Early British Computers, Manchester University, Manchester, 1980
  • Raúl Rojas, Ulf Hashagen (editors), The First Computers — History and Architectures, MIT Press, Cambridge, 2000 - the Mark I is covered in Part IV-2, pp. 369-377
  • B. V. Bowden (editor), Faster Than Thought: A Symposium on Digital Computing Machines, Pitman Publishing, New York, 1964 - the Mark I is covered in Chapter 6, pp. 117-124; written by Tom Kilburn

External links