COLLOQUIUM
Department of Computer Science and Engineering
University of South Carolina
 
Dynamic Load Balancing 
for Parallel and Distributed Systems 

Zhiling Lan*
Northwestern University 

Date: February 8, 2002 (Friday)
Time:  3:30-4:30PM
Place: Swearingen 2A27

Abstract

Adaptive Mesh Refinement (AMR) is a type of multiscale algorithm that
achieves high resolution in localized regions of dynamic,
multidimensional numerical simulations. One of the key issues related
to AMR is dynamic load balancing (DLB), which allows large- scale
adaptive applications to run efficiently on parallel and distributed
systems. In this talk, I will first provide a detailed adaptive
characterization of AMR applications, then two schemes are proposed to
efficiently redistribute workload among homogeneous and heterogeneous
systems respectively. For the parallel systems, the proposed scheme
integrates a grid-splitting technique with direct grid movements. For
the distributed systems, both the heterogeneous and dyanmic features
of distributed systems are addressed and a heuristic method is
presented to evaluate the computational gain and redistribution cost
for global redistribution. Experiments show that by using the proposed
DLB schems, the execution time can be reduced by up to 57% and the
quality of load balancing can be improved by a factor of four.

Zhiling Lan is currently a Ph.D candidate at the ECE department of
Northwestern University, and is expected to receive her Ph.D degree in
June, 2002. Her research interests include parallel and distributed
systems, and performance modeling and analysis.  She received a BS in
Mathematics from Beijing Normal University and an MS in Applied
Mathematics from Chinese Academy of Sciences.


*Ms. Lan is a candidate for a faculty position in the Department of
Computer Science and Engineering, University of South
Carolina---Columbia.