1900 Series computers

These varied in computing power and the models included:
ICT 1901
ICT 1902
ICT 1903
ICT 1904
ICT 1905 (similar to the 1904 but incorporating an autonomous floating-point unit)
ICT 1906
ICT 1907 (similar to the 1906 but incorporating an autonomous floating-point unit)
ICT 1909 (similar to the 1905 but with low-cost 6µs core store)
The execution time for an addition instruction ("add the contents of store location x to register y") ranged from 2.5μs for a 1906 or 1907 with 1.1μs core store, to 34μs for a 1901 with 6μs core store.
The first machine was the 1904 which was a Canadian design from the Ferranti-Packard company, originally called the FP6000, although commonly referred to within West Gorton as the FP1. It is said that this machine with core store memory fired up with its program still in store after its sea freight journey from Canada.[citation needed]
One feature of these mainframes was the common instruction set throughout the range, meaning that a program written and compiled on one machine would run unchanged on any larger machine in the range, and also on any smaller machine that had all the facilities used by the program. In fact the hardware was different between machines. To achieve this a program termed "the executive" or exec encapsulated the hardware and supplied software routines (extracodes) to supplement the hardware supplied instructions.
By 1968 ICT had merged with English Electric computers and become ICL.
Enhanced versions of the 1900 series subsequently appeared with an A suffix, an E suffix, an F suffix (for floating point), a T suffix, and an S suffix, e.g.
The operation in Stevenage, Hertfordshire, designed and built machines, typically smaller systems using the E3RM executive software
ICL 1901A
ICL 1901S/T
ICL 1902A
ICL 1902S
ICL 1902T
ICL 1903A
ICL 1903S
The operation at West Gorton, Manchester, built machines, typically larger systems using the E6RM executive software:
ICL 1903T
ICL 1904A
ICL 1904T
ICL 1904S
ICL 1905F
ICL 1906S
The 1900 Series were 24-bit word machines, supporting four 6-bit characters per word and using octal for binary short-hand — as opposed to the IBM systems, which used 8-bit bytes and hex. Basic memory on the smaller machines was 16K words (or 64 kilobytes equivalent), and there were even 8K word versions sold (although most actually had 16K memory factory installed in readiness for the certain upgrade order that followed!)—early machines using core memory (ferrite rings on a copper wire matrix) and operated on binary hand-switches on the mainframe cabinet. Despite the apparent small memory size, quite sophisticated applications were run on the equipment and computer programmers paid great attention to the efficient use (and reuse) of memory.
Input and output (I/O) consisted of 80-column cards (40-column cards—with round holes—were unable to cope with the full 64-character set), a card punch and 8-track paper tape; printed output was produced on a solid barrel line printer (120 columns wide).
The first commercial sale was made in 1964 to the Morgan Crucible Company and consisted of a 16K word 1902 with an 80-column 980-card/minute reader, a card punch, a 600 line/min printer and 4 x 20kb/s tape drives. It was soon upgraded to a 32K word memory and a floating point unit to allow for some scientific work. Incidentally this company was also the first business to order ICT's previous computer, the HEC4 (later ICT 1201) in 1955.
Early machines used storage on reels of half-inch magnetic tape and were then augmented by direct access devices (EDS, Exchangeable Disc Store) typically with capacities of 1.6Mch, 4Mch and 8Mch. By the time 30Mch packs were available they occupied a cabinet 4 feet (1.2 m) high.
After ICL had introduced the "New Range" (NR) series of computers during the 70s, marketed as the 2900 series, the 1900 series actually lived on as there was such a wealth of software written for it:
The ICL 2903/2904, designed and built in Stevenage, which were really successors to the 1902A/3A machines as they had the same target instruction set (albeit with a native 32 bit rather than 24 bit architecture)
Second generation "S3E" (microcoded) versions of the larger NR systems (such as the 2960/2966 from West Gorton, and the later 2940/50 from Stevenage), could run 1900 series code under DME (Direct Machine Environment) as an emulation as well as the New Range instruction set under the newer VME (Virtual Machine Environment)
The later availability of CME (Concurrent Machine Environment) microcode, which allowed DME and VME to co-exist (and run) concurrently on the same platform, similar to the functionality offered by virtualisation software such as VMware today
The replacement for the 2903/4 was the ME29 (Machine Environment 29) with its own operating system, TME (Transaction Machine Environment), which could still execute 1900 applications
An update of CME to CME* in the 1980s to support a small VME range called DM1 that replaced the ME29, which could run TME within a VME VM (virtual machine)
Programming languages included the assembly type language PLAN (Programming LAnguage Nineteen-hundred) and latterly COBOL for the development of commercial-orientated systems while ALGOL and FORTRAN were used for scientific work.
The basic operating system was called "Executive" and supported "multiprogramming", i.e. it was capable of running up to four programs concurrently on the 1904, and more on the larger machines. From about 1968 onwards, advanced features such as a tree-structured filestore, batch processing and multi-access with identical command languages, and spooling were provided by a privileged ("trusted") program called GEORGE.