Dynamic Statistics of Sequential Prolog

Ignacio Celis and Jonathan W. Mills
Computer Science Dept.
Indiana University
Lindley Hall 215
Bloomington, IN 47405
e-mail: {celisi, jwmills}@cs.indiana.edu
phone (812)-857-7038

Keywords: Implementations and Architectures

Abstract
This paper presents dynamic statistics taken by simulating sequential Prolog programs. We used the Warren's Abstract Machine (WAM) as the abstract machine to simulate these programs. We show the dynamic data for the following areas: Dereferencing Chain, Unification, Trailing, Untrailing, Choice Point Creation, Environment Creation, Backtracking, etc. This information can be useful for compiler writers and special purpose hardware designers to concentrate their effort toward those features of Prolog that are more frequently used.