The Community for Technology Leaders
16th Euromicro Conference on Parallel, Distributed and Network-Based Processing (PDP 2008) (2014)
Torino, Italy Italy
Feb. 12, 2014 to Feb. 14, 2014
ISSN: 1066-6192
pp: 228-231
Ivanoe De Falco , Inst. of High-Performance Comput. & Networking, Naples, Italy
Eryk Laskowski , Inst. of Comput. Sci., Warsaw, Poland
Richard Olejnik , Comput. Sci. Lab. of Lille, Univ. of Sci. & Technol. of Lille, Lille, France
Umberto Scafuri , Inst. of High-Performance Comput. & Networking, Naples, Italy
Ernesto Tarantino , Inst. of High-Performance Comput. & Networking, Naples, Italy
Marek Tudruj , Inst. of Comput. Sci., Warsaw, Poland
ABSTRACT
The paper concerns methods for using Extremal Optimization (EO) for processor load balancing during execution of distributed programs. A load balancing algorithm for clusters of multicore processors is presented and discussed. In this algorithm the EO approach is used to periodically detect the best tasks as candidates for migration and for a guided selection of the best processors to receive the migrated tasks. To decrease the complexity of selection for migration, we propose a guided EO algorithm which assumes a two step stochastic selection during the solution improvement based on two separate fitness functions. The functions are based on specific program models which estimate relations between the programs and the executive hardware. The proposed load balancing algorithm is assessed by experiments with simulated load balancing of distributed program graphs. The algorithm is compared against an EO - based algorithm with random placement of migrated tasks and a classic genetic algorithm.
INDEX TERMS
Load management, Genetic algorithms, Heuristic algorithms, Optimization, Measurement, Monitoring, Instruction sets
CITATION

I. De Falco, E. Laskowski, R. Olejnik, U. Scafuri, E. Tarantino and M. Tudruj, "Extremal Optimization with Guided State Changes in Load Balancing of Distributed Programs," 16th Euromicro Conference on Parallel, Distributed and Network-Based Processing (PDP 2008)(PDP), Torino, Italy Italy, 2014, pp. 228-231.
doi:10.1109/PDP.2014.56
183 ms
(Ver 3.3 (11022016))