Implementation of a pattern-matching approach for identifying algorithmic concepts in scientific FORTRAN programs

摘要

A significant barrier inhibiting the use of parallel computing is the difficulty of writing parallel software. One solution to this problem is to build tools that can automatically convert sequential programs to parallel programs. In this paper we describe the implementation of a system that is designed to perform a semantic level analysis and transformation of algorithms in sequential programs. Our approach is based on pattern matching and has been inspired by recent research on reverse engineering and program understanding. We describe the architecture of our system, a pattern matching language that we have designed, our pattern matching strategy, and illustrate the approach and implementation using examples. This paper is a summary of the work.