Bus versus cellular automata, and ultimate limitations of parallel processing

  • Authors:
  • Jerome Rothstein

  • Affiliations:
  • Department of Computer and Information Science, The Ohio State University, 2036 Neil Avenue Mall, Columbus, OH 43210-1277 USA

  • Venue:
  • Mathematical and Computer Modelling: An International Journal
  • Year:
  • 1988

Quantified Score

Hi-index 0.98

Visualization

Abstract

After a general survey of current limitations on computation and data processing, we concentrate on that of raising speed by parallel/distributed methods. The major conceptual problem is how to make effective use of large numbers of potentially available modules or subsystems. This is accomplished in bus automata, by combining the parallel processing capacity of cellular automata with the communicative power of iterative switching arrays. Each cell controls its portion of the array, whereby cooperating cells can establish program-controlled communication-bus patterns which permit direct communication between separated groups of cells. Up to theoretically maximum useable amounts of parallel capacity can be effectively marshalled to solve many complex problems ''immediately.'' Others are sped up enormously even if immediacy is not obtained. Results to date on immediate computation, language parsing, and pattern recognition will be reviewed and open questions discussed. Applications are indicated to problems which nervous systems generally, and visual systems in particular handle rapidly and in parallel. The bus automaton is suggested for the next generation of brain and mental models.