Priming the Pump: Load Balancing Iterative Algorithms

Abstract

Load balancing iterative algorithms is an interesting problem in resource allocation that is useful for reducing total elapsed processing time through parallel processing. Load balancing means that each processor in a parallel processing environment will handle about the same computational load. It is not sufficient to allocate the same number of processes to each processor since different processes or tasks can require different loads . For iterative algorithms, load balancing is the process of distributing the iterations of a loop to individual processes . This paper will analyze different methods used for load balancing. Each method will be measured by how well it reduces the total elapsed time and by algorithm complexity and overhead. Measured data for different load balancing methods will be included in this paper.