A piggy-back compiler for Prolog

In this paper, we describe’ the design and impienieutation of an optimizing Prolog compiler. Our approach is novel both for the breadth of techniques employed and the performance we achieve as a resuit of using those techniques. Furthermore, we know of no other compiler for Prolog that utilizes the same techniques and achieves comparable performance. h addition to efficiency, the design goals were portah’lity and simplicity. Our compiler takes advantage of knowledge about what it is compiiing to reduce the mechanisms needed to support its runtime evaluation. The types of knowledge it uses are: