As well as the 1900 series machine, ICL had a range of other
systems, some inherited through the various mergers that
formed ICL and others developed as replacements or as new
systems.
| System |
Year |
Brief Details |
| PEGASUS |
1956 |
The Ferranti PEGASUS was a vacuum-tube machine with
a word length of 39 bits. |
| 1101 |
1960 |
The ICT 1101 was the badged EMI (Electric and Musical
Industries) EMIDEC 1100. |
| 1201 |
|
The BTM 1201
System was a hybrid with a tabulator. |
| 803 |
1962 |
This machine was produced by Elliot Brothers and was
built using discrete transistor and ferrite-core
technology, and using bit-serial logic. The word length
was 39 bits. (See
Bill Purvis's Site or
Dave Brooks's Site)
Note: Both links are now defunct. |
| 1004 |
1962 |
ICT agreement with Univac to sell the 1004. |
| 1301 |
1962 |
The ICT 1301
System came from a design with GEC. |
| 1500 |
1962 |
This was a badged RCA 301. |
| ATLAS |
1962 |
The Ferranti ATLAS came from work with Manchester
University (Mark 1) and became the ICT Atlas, these
machines ran until the early 1970s.
When the first one was delivered in December 1962,
it was considered to be the fastest computer in the
world.
Job Logs
from an Atlas system |
| 1900 |
1964 |
The Ferranti-Packard 6000 became the ICT (later ICL)
1900 series, as a direct response to IBM's
announcement of the 360 series. The 1900 series was
announced about six months later, but gained ground
in the UK and Europe as they were cheaper and early
machines were delivered before the 360s. (See
Marketing
Newsletters.) This series of machines were still
running into the 1980s, and the software continued
on the 2900 series under DME. |
| System/4 |
|
The System/4 came from English Electric and was an
IBM 360 clone. The System/4-72 was the RCA Spectra
70, made under license. |
| System 10 |
|
This came from Singer, a US Company, which was bought
by ICL, I assume in an attempt to get into the US
market.
See:-
http://members.iinet.net.au/~daveb/Sys-10.html
Note: Link extinct. |
| 2903 |
1973 |
This was a series of small mainframes (2903, 2904 &
2905) which were really 1900s built using 2900
technology and in orange (the 2900 colour scheme)
boxes. |
| 2900 |
1975 |
This was the awaited 'New Range' system, which
replaced the 1900 and System/4, together with a new
operating system VME (Virtual Machine Environment).
The machines could still run as 1900 systems using
the DME microcode option.
It was rumoured, when I
worked for ICL, that the DME microcode was 'slugged'
to make the VME systems look better, and GEORGE 3
running on 'full speed' DME considerably outperformed
the 'native' mode VME.
See:-
Early Computing at Edinburgh University |
| ME29 |
1980 |
The ME29 was a mid-range mainframe, originally
a larger version of the 2903 series, later gaining a
new 2900 style operating system TME, but still
running 1900 software. |
| System 25 |
1981 |
The System 25 was the replacement for the System 10
|
| PERQ |
|
This was a Graphics Workstation, licenced from the 3
Rivers Computer Corporation and ran under POS (Perq
Operating System) or PNX (Perq Unix). They were
P-Code machines and had user enhanceable microcode.
Pascal (the 'assembler') and Fortran 77 were
available under POS. They had a Keyboard, A4 black
and white screen, serial and GPIB ports and used a
graphics tablet (connected via GPIB) rather than a
mouse.
These machines were my introduction to networking, I
was give a large box of bits (circuit cards, drop
cables,transceivers, thick ethernet cable,
terminators and notes) and told to connect our
departments machines together in a network. It worked,
well sort of, at that the time the software was a bit
flakey but the hardware side seemed okay.
We weren't allowed to call it Ethernet, it was OS/LAN
(Open Systems Local Area Network) and intended to
allow ALL ICL systems to interconnect, from the
largest 2900 mainframe down to a terminal. |
| Personal Computer |
|
The first range of PCs were based on the 8-bit Intel
8085 microprocessor, running CP/M and on the larger
model MP/M. The machines had 64K RAM (256K banked as
4 x 64K for the MP/M system), 4 serial ports, 1 x
5.25" floppy drive and either a 2nd floppy drive or a
hard drive (5 or 10Mb). The MP/M model supported up
to 3 users (VDUs) and 1 printer, but could be
expanded with a 2nd memory card and port extension
unit.
The later 16-bit machines ran CCP/M, later Concurrent
DOS, and were all multi-user, allowing up to 4 VDUs
and 2 printers as standard. The sales of these
machines dropped when the single user IBM PC came
onto the market. |
| DRS20 |
1981 |
The DRS20 range of machines was based on the 8-bit
8085 microprocessor and were a network system, based
around a LAN (2MB on RG68 cable). The machines had
multi-processors, depending on the model/function.
A standard workstation (model /10 or /110) would have
2 processors, a control processor which ran the
operating system and also controlled the keyboard and
screen, and an application processor which ran the
user program.
A workstation/file server (e.g. model /150) would
have 4 processors, a control processor which ran the
operating system (plus keyboard and screen), a file
processor, a network processor and an applications
processor as standard. Additional processor boards
could be fitted, e.g. a communications processor (WAN
link, automated Telex), a second application
processor (print spooler).
The system was very flexible on configuration and
included Cobol (MF CIS Cobol) and Pascal compilers, a
word processor, FTP and remote terminal emulation
software to link to an ICL mainframe.
There was a later /2xx series of machines which had
16-bit application processors. |
| DRS300 |
|
The DRS300 range replaced both the DRS20 and the PC.
It was a modular system, based on units about the
size of an A4 box file, which clipped together to
build the required configuration. The original
systems ran Concurrent DOS and then a Unix version
was introduced. The terminals connected via a LAN.
|
| Series 39 |
1985 |
The Series 39 was the replacement for the 2900 series
machines. |