Performance analysis of garbage collection and dynamic reordering in a Lisp system

Generation-based garbage collection and dynamic reordering of objects are two techniques for improving the efficiency of memory management in Lisp and similar dynamic language systems. An analysis of the effect of generation configuration is presented, focusing on the effect of the number of generations and generation capacities. Analytic timing and survival models are used to represent garbage collection runtime and to derive structural results on its behavior. The survival model provides bounds on the age of objects surviving a garbage collection at a particular level. Empirical results show that execution time is most sensitive to the capacity of the youngest generation. The existence of a range of optimum values demonstrates the potential for the tuning of garbage collection.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Electrical Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2011

Author dc:creator

• Llames, Rene Lim

Contributors dc:contributor

• Iyer, Ravishankar K.

• Computer Science

Statement dc:rights

• Copyright 1991 Llames, Rene Lim

Language dc:language

eng

Identifier

AAI9136665
(UMI)AAI9136665

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/19302