A Novel Scheduling Algorithm Based on Critical Path in Distributed Heterogeneous Computing Systems

초록

Efficient DAG scheduling is critical for achieving high performance in heterogeneous computing systems. Finding an optimal solution to the problem of scheduling an application modeled by a directed acyclic graph(DAG) onto a set of heterogeneous machines is known to be an NP-complete problem. In this paper we propose a new list scheduling algorithm, called the Critical Path Rank Scheduling(CPRS),to exploit all of a program’s available parallelism in distributed heterogeneous computing system. The primary goal of CPRS is to minimize the schedule length of applications. The performance of the algorithm has been observed by its application to some practical DAGs, and by comparing it with other existing scheduling algorithm in term of the schedule length. The comparison studies show that CPRS significantly outperform PETS, HCPT and CPOP in terms of performance.