Fifth generation computer systems will represent a unification of research into VLSI processors and into distributed processing. Each computer system will consist of a network of computing elements supporting an individual application or need. VLSI processor research will allow a computing element to provide either a general-purpose or a special-purpose function, and range in power from a main-frame computer to a miniature microcomputer. Distributed processing research will allow a network to be physically dispersed across a country or building, or to be physically close as on a single highly integrated chip. For this spectrum of fifth generation computer systems to be programmed as individual computers and for their computing elements to work together, it is necessary to have a single computer architecture to which they all conform. The basis of this fifth generation computer architecture could be control flow, data flow or reduction computers. Here I examine these three types of computer and present classifications for their underlying concepts, with the aim of analysing their probable contribution to fifth generation computer architecture.
[1]
David A. Patterson,et al.
X-Tree: A tree structured multi-processor computer architecture
,
1978,
ISCA '78.
[2]
Gerald Jay Sussman,et al.
Scheme-79 - Lisp on a Chip
,
1981,
Computer.
[3]
Edward P. Farrell,et al.
A concurrent computer architecture and a ring based implementation
,
1979,
ISCA '79.
[4]
A. L. Davis,et al.
The architecture and system method of DDM1: A recursively structured Data Driven Machine
,
1978,
ISCA '78.
[5]
H. T. Kung,et al.
The Design of Special-Purpose VLSI Chips
,
1980,
Computer.
[6]
Robert M. Keller,et al.
A loosely-coupled applicative multi-processing system*
,
1979,
1979 International Workshop on Managing Requirements Knowledge (MARK).