Abstract

Scientific Programming and subsequent compilation is significantly complicated when programs are expected to perform on one or many processors for any size or dimensional problem. Scientific languages, although sophisticated and powerful, have been slow to evolve from their Von Neumann scalar by scalar operations to monolithic data parallel functions, e.g. Fortran 90/95. The structure of an array and corresponding architectural topology to which the arrays are mapped, may provide ways to increase portability and scalability of designs. This paper presents how the Psi(w) Calculus in conjunction with a useful algebra that provides convolutions, permutations, and scalar operations, is being used to aid scientists' thinking and programming in the monolithic structures common to their disciplines, while allowing them continued use of the programming languages they've learned and enjoyed.

Department(s)

Computer Science

Comments

The second Author is a Graduate Student.

This work was supported by the National Science Foundations's PFF Award.

Keywords and Phrases

Scientific Computation, Array Languages, HPF, Program Verification, Church-Rosser Property, Reduction Rules, Denotational Semantics, Operational Semantics, Portable and Scalable Computing.

Report Number

CSc-94-29

Document Type

Technical Report

Document Version

Final Version

File Type

text

Language(s)

English

Rights

© 1994 University of Missouri - Rolla, All rights reserved

Publication Date

12 January, 1995

Download

Share

 
COinS