Task Mapping Algorithm for Heterogeneous Computing System Allowing High Throughput and Load Balancing

The applicability and strength of heterogeneous computing systems are derived from their ability to match computing nodes to appropriate tasks since a suite of different machines are interconnected. A good mapping algorithm offers minimal expected completion time and machine idle time. In this paper we propose a new task scheduling algorithm allowing higher performance than the existing algorithms such as the Min-min, Max-min, and Sufferage algorithm. It is achieved by task swapping approach based on the expected completion time and ready time of each machine. Extensive computer simulation validates the proposed algorithm along with the earlier ones.

[1]  R. F. Freund,et al.  Scheduling resources in multi-user, heterogeneous, computing environments with SmartNet , 1998, Proceedings Seventh Heterogeneous Computing Workshop (HCW'98).

[2]  Howard Jay Siegel,et al.  Representing Task and Machine Heterogeneities for Heterogeneous Computing Systems , 2000 .

[3]  Min-You Wu,et al.  A high-performance mapping algorithm for heterogeneous computing systems , 2001, Proceedings 15th International Parallel and Distributed Processing Symposium. IPDPS 2001.

[4]  Albert Y. Zomaya Parallel and Distributed Computing Handbook , 1995 .

[5]  Howard Jay Siegel,et al.  Task execution time modeling for heterogeneous computing systems , 2000, Proceedings 9th Heterogeneous Computing Workshop (HCW 2000) (Cat. No.PR00556).

[6]  Yves Robert,et al.  Scheduling tasks sharing files on heterogeneous master-slave platforms , 2004, 12th Euromicro Conference on Parallel, Distributed and Network-Based Processing, 2004. Proceedings..

[7]  Oscar H. Ibarra,et al.  Heuristic Algorithms for Scheduling Independent Tasks on Nonidentical Processors , 1977, JACM.

[8]  R. F. Freund,et al.  Dynamic matching and scheduling of a class of independent tasks onto heterogeneous computing systems , 1999, Proceedings. Eighth Heterogeneous Computing Workshop (HCW'99).