Workload Distribution on Heterogeneous Platforms

This paper targets the problem of finding an efficient distribution of a computational task on a heterogeneous computing platform. The heterogeneity of the processing elements arise due to differences in computation speed and memory capacity of the processors. We first consider using a discrete functional performance model that integrates processing speed and capacity of processing elements and then develop a mathematical model and propose a heuristic mapping algorithm for distributing a given total workload of size N on p processing elements such that the total computation time is minimized. Computational results show that the proposed method provides a significant improvement in reducing the computation time in comparison to equal distribution approach.