Implementation of a high-speed Prolog interpreter

Abstract

This paper describes the implementation of a high speed Prolog interpreter on a standard microprocessor (50 KLIPS on a 16 MHz MC68020). The interpreter is based on direct threaded code. By this method an interpreted program achieves the same speed as a compiled program, but uses only a tenth of memory. The first part of this paper describes the implementation of the interpreter. The second part compares the implementation, the runtime and the storage requirements with that of a compiler.