Dynamic Load Distribution Optimization in Heterogeneous MultipleProcessor Systems

Abstract

We examine the problem of optimizing the distribution of the m interacting modules of a given work load on a parallel system with p heterogeneous processors. Average-valued parameters are used to model the intermodule coupling of the work load and its execution and communication times on the diverse system processors. We derive an analytical optimality criterion to minimize a multi-metric objective function representing a weighted combination of work load completion time, communication cost, resource utilization cost, and processor idle-time.