LI-Engine: a New Sequential Control Model for Prolog

Exploitation of architectural support for Prolog-like languages takes one of two main paths: compiler-based or interpreter-based. The former scheme enforces the transformation from a logic program to a set of instructions tailored for Prolog to accommodate the instruction-driven von Neumann processor, while the latter implements a complicated inference procedure as a microinterpreter to realize the relation-driven paradigm of logic programming. In this paper, we present a new sequential control model developed in the design of the LI-Engine. The control unit of LI-Engine is designed by fully exploiting the simplicity of the relation-driven paradigm. The proposed algorithm adopts the post-choicemethod and involves only four basic cache operations, ten assignments and four iterators controlled by simple Boolean tests. We believe that this research o ers another possible direction for developing high performance logic inference machines.