The impact of the LCC-optimization on the performance of a static scope LISP interpreter

Abstract

The Low Cost Call - Optimization was introduced as a general approach to the optimization of function calls in shallow binding implementations of applicative static scope languages. The basic idea of the LCC-optimization is to avoid saving obsolete environments for a large class of function calls, thus exceeding the usual optimization of tailrecursive and covered tailrecursive calls. This paper describes how the new technique has been applied in the implementation of a portable static scope LISP interpreter for microcomputers, called LCINT. In a benchmark test we measure the impact of the LCC-optimization on the performance of LCINT. The results show that this technique is particularly valuable for small systems with a rather limited storage. Whereas some of the results fully meet our expectations, others are rather surprising and lead us to certain conclusions for further research.