The Bendix G-20
computer was introduced in 1961 by the Bendix Corporation
, Computer Division, Los Angeles, California. The G-20 followed the highly successful G-15 vacuum tube computer. Bendix sold its computer division to Control Data Corporation
in 1963, effectively terminating the G-20.
The G-20 system was a general purpose mainframe computer
, constructed of transistorized modules and core memory. Memory size was 32 bits, plus parity. Up to 32k words of memory could be used. Single and Double precision floating point were allowed, as well a custom scaled format, called Pick-a-Point. A special form of the pick-a-point allowed an integer.
Memory locations 1 through 63 were used as index registers. One hundred and ten instructions were in the instruction set. The CPU included integral block I/O and interrupt facilities. Multiply time was 51-63 microseconds, and divide time was 72-84 microseconds. Basic memory cycle time was 6 microseconds.
A special configuration of the G-20, a dual processor G-21, was used to support campus computing at Carnegie Institute of Technology
in the 1960s. Usually the two processors ran independently, one CPU handling card-based input, and the other handling jobs submitted through one of 16 AT&T Dataphones connected to telephone lines, usually via Teletype
35 KSR, 35 ASR and Model 33 ASR
teleprinters. The G-21 had 32k words of memory for each processor, but could be reconfigured for 64k mode for large programs, usually as a... Read More