Characterizing the paralation model using dynamic assignment
Collection-oriented languages provide high-level constructs for describing computations over collections. These languages are becoming increasingly popular with the advent of massively parallel SIMD machines. Even on serial machines, collection-oriented languages are interesting because they allow the user to describe computation in a concise manner.
In this paper we implement Sabot's paralation model , a collection-oriented language extension for expressing elementwise computation and explicit communication. We extend the Scheme programming language, a dialect of Lisp with lexical scope and first-class procedures, basing our implementation on Sabot and Blelloch's original code written in Common Lisp. We have taken particular care to remove all unnecessary side-effects. This re-packaging has taken two forms. First, many of the side-effects can be treated as a kind of dynamic assignment. Some others can be removed by coding in a more functional style. The remaining category of side-effects are for site-specific operations. This re-packaging has led to a surprising result: the paralation model can be described with just a single language form and a single primitive.
KeywordsFormal Parameter Closed Mapping Target Field Abstract Data Type Index Field
Unable to display preview. Download preview PDF.
- 1.Guy E. Blelloch: Vector Models for Data-Parallel Computing. The MIT Press, Cambridge, Massachusetts, 1990.Google Scholar
- 2.Guy E. Blelloch and Gary W. Sabot: Compiling Collection-Oriented Languages Onto Massively Parallel Computers. J. of Par. and Distr. Comp. 8, 2, 119–134.Google Scholar
- 3.C. J. Date: An Introduction to Database Systems. Addison-Wesley, Reading, Massachusetts, 1977.Google Scholar
- 4.R. Kent Dybvig: The Scheme Programming Language. Prentice-Hall, Englewood Cliffs, New Jersey, 1987.Google Scholar
- 5.Daniel P. Friedman, Mitchell Wand, and Christopher T. Haynes: Essentials of Programming Languages. The MIT Press, Cambridge, Massachusetts, 1992.Google Scholar
- 6.Daniel W. Hillis, and Guy L. Steele Jr.: Data Parallel Algorithms. Communications of the ACM 29, 12, 1170–1183.Google Scholar
- 7.Gary W. Sabot: The Paralation Model: Architecture-Independent Parallel Programming. The MIT Press, Cambridge, Massachusetts, 1988.Google Scholar
- 8.Guy L. Steele Jr.: Common LISP: The Language Second Edition. Digital Press, 1990.Google Scholar