Problem formulation using array processing techniques

  • Authors:
  • D. J. Evans

  • Affiliations:
  • University of Technology, Loughborough, U.K.

  • Venue:
  • Proceedings of the conference on Programming languages and compilers for parallel and vector machines
  • Year:
  • 1975

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper outlines the evolution of array processing computational techniques and systems and briefly discusses the important concepts of the software and programming strategies which enable these systems to be used effectively. For very large problems, the advantages of parallel processing techniques are stressed and the use of a mathematical programming array language outlined for removing some of the arbitrariness from the problem formulation and programming stages of solving partial differential equations. The purpose of the proposed approach is to eliminate errors in the logical formulation early in the design of a large scientific program and hence prevent many of the logical errors that can creep into a program because of its sheer size and complexity. The discussion includes the solution of both boundary and initial value scientific problems using an approach in which APL augments FORTRAN in the problem formulation stage.