COLLOQUIUM
Department of Computer Science and Engineering
University of South Carolina
 
Statistical Learning Techniques for
Intelligent Memory Management

Nancy Glenn

Department of Statistics
University of South Carolina

Date: April 25, 2003 (Friday)
Time: 3:30-4:30PM
Place: Swearingen 1A03 (Faculty Lounge)

Abstract

We explore the use of statistical properties of the directed graph
describing the set of live data-data reachable from a set of roots.
We use these properties to decide between garbage collection and heap
expansion in a memory management algorithm that combines heaps
represented by dynamic arrays with a mark-and-sweep garbage collector.
Our goal is to improve overall execution performance significantly.
The sampling method that predicts the density of useful data is
implemented as a partial marking algorithm, and its computation time
is reusable by the garbage collection algorithm.  The resulting
heuristics are tested empirically in the context of the Jinni Prolog
compiler's runtime system.

Nancy Glenn obtained a B.S. in Mathematics, and a B.S. in Statistics
from the University of South Carolina.  She obtained her Ph. D. in
2001 from Rice University's Department of Statistics, which is in the
School of Engineering.  Her areas of expertise are Data Mining and
Nonparametric Statistics.  She is a member of the American Statistical
Association, and the Society of Industrial and Applied Mathematics.
She is currently an Assistant Professor in USC's Department of
Statistics.