Fast algorithms for fair dynamic load redistribution in heterogeneous environments

Authors:
Tiberiu Rotaru;Hans-Heinrich Nägeli
Affiliations:
Fraunhofer-Institut für Techno-und Wirtschaftsmathematik, Gottlieb-Daimler-Str. 49, 67663 Kaiserslautern, Germany;Institut d'informatique, Université de Neuchâtel, Émile-Argand 11, Case Postale 2, CH-2007 Neuchâtel, Switzerland
Venue:
Applied Numerical Mathematics - Numerical algorithms, parallelism and applications
Year:
2004

Citing 10
Cited 0

Parallel and distributed computation: numerical methods

Parallel and distributed computation: numerical methods
Dynamic load balancing for distributed memory multiprocessors

Journal of Parallel and Distributed Computing
Load balancing and Poisson equation in a graph

Concurrency: Practice and Experience
Handbook of combinatorics (vol. 1)

Handbook of combinatorics (vol. 1)
Matrix computations (3rd ed.)

Matrix computations (3rd ed.)
An improved diffusion algorithm for dynamic load balancing

Parallel Computing
Efficient schemes for nearest neighbor load balancing

Parallel Computing - Special issue on parallelization techniques for numerical modelling
Matrix analysis and applied linear algebra

Matrix analysis and applied linear algebra
Load Balancing in Parallel Computers: Theory and Practice

Load Balancing in Parallel Computers: Theory and Practice
Heterogeneous Dynamic Load Balancing

IWCC '01 Proceedings of the NATO Advanced Research Workshop on Advanced Environments, Tools, and Applications for Cluster Computing-Revised Papers

Quantified Score

Hi-index	0.00

Visualization

Abstract

Efficiently running parallel applications on distributed computing environments requires adequate fair dynamic load redistribution techniques. In this paper, two fast algorithms for maintaining the fairness in a dynamic heterogeneous context are discussed. The first algorithm is based upon a conjugate gradient method and solves a quadratic optimization problem. The second algorithm finds the solution of the same problem using the characteristic polynomial of the generalized Laplacian of the communication graph. Both algorithms are faster than the well-known diffusion algorithms. Experimental results are reported.