Data structures for associative supercomputers

Author

Potter, J.L.

Author_Institution

Kent State Univ., OH, USA

fYear

1988

fDate

10-12 Oct 1988

Firstpage

77

Lastpage

84

Abstract

The simplicity of programming associative supercomputers is described, and conventional associative data references are expanded into a generalized associative data structure reference mechanism which encompasses the arrays, data structures, and data-typing constructs of conventional languages. A unified approach is presented for representing arbitrarily complex data structures in content addressable memories and associative computers. This approach to data structures in associative computers has the advantages of (1) automatically extracting fine-grain parallelism. (2) eliminating much of the complexity of the nonalgorithmic address computation in program development, (3) allowing multiple data structures to be associated with each datum, (4) allowing the data structures themselves to be modified, and (5) allowing information exchange between vastly different program languages such as Lisp, Prolog, OPS5, Fortran and Pascal

Keywords

computational complexity; content-addressable storage; data structures; parallel processing; Fortran; Lisp; OPS5; Pascal; Prolog; arrays; associative supercomputers; complexity; content addressable memories; data structures; data-typing constructs; fine-grain parallelism; nonalgorithmic address computation; programming; reference mechanism; Associative memory; Broadcasting; Cams; Central Processing Unit; Computational modeling; Concurrent computing; Data structures; Parallel processing; Read-write memory; Supercomputers;

fLanguage

English

Publisher

ieee

Conference_Titel

Frontiers of Massively Parallel Computation, 1988. Proceedings., 2nd Symposium on the Frontiers of

Conference_Location

Fairfax, VA

Print_ISBN

0-8186-5892-4

Type

conf

DOI

10.1109/FMPC.1988.47416

Filename

47416