Comparative Analysis of Classic Garbage-Collection Algorithms for a Lisp-like Language

TL;DR

This paper compares several garbage collection algorithms for Lisp-like languages, demonstrating Cheney's Copy Algorithm as particularly effective and analyzing the challenges of developing an optimal collector.

Contribution

It provides a comparative analysis of classic garbage collection algorithms in a Lisp environment, highlighting Cheney's algorithm's advantages and discussing core design considerations.

Findings

Cheney's algorithm outperforms Mark-Sweep in Lisp-like systems

Optimal garbage collector development is infeasible for general use

Experimental results favor Cheney's algorithm for Lisp environments

Abstract

In this paper, we demonstrate the effectiveness of Cheney's Copy Algorithm for a Lisp-like system and experimentally show the infeasability of developing an optimal garbage collector for general use. We summarize and compare several garbage-collection algorithms including Cheney's Algorithm, the canonical Mark and Sweep Algorithm, and Knuth's Classical Lisp 2 Algorithm. We implement and analyze these three algorithms in the context of a custom MicroLisp environment. We conclude and present the core considerations behind the development of a garbage collector---specifically for Lisp---and make an attempt to investigate these issues in depth. We also discuss experimental results that imply the effectiveness of Cheney's algorithm over Mark-Sweep for Lisp-like languages.