Scheduling parallel tasks onto numa multiprocessors with inter-processor communication overhead

0Citations
Citations of this article
1Readers
Mendeley users who have this article in their library.
Get full text

Abstract

This chapter addresses the problem of scheduling tasks in the Non-Uniform Memory Access (NUMA) multiprocessor system with a bounded number of available processors. An algorithm is proposed here to schedule tasks by considering the intertask communication overhead and the contentions among communication channels. Communication contentions arise from the communication medium having insufficient capacity to serve all transmissions, causing significant contention delays. The proposed algorithm also exploits the schedule-holes in schedules; therefore, it could produce better schedules than that produced by existing algorithms. In this chapter, a sharper bound in the multiprocessor scheduling problem with the consideration of the communication delay is also shown. The proposed algorithm ensures performance within a factor of two times of the optimum scheduled lengths for general directed acyclic task graphs. Experimental results demonstrate the superiority of the proposed algorithm over that presented in literature. © 2005 Springer Science+Business Media, Inc.

Cite

CITATION STYLE

APA

Lai, G. J. (2005). Scheduling parallel tasks onto numa multiprocessors with inter-processor communication overhead. In New Horizons of Parallel and Distributed Computing (pp. 97–112). Springer US. https://doi.org/10.1007/0-387-28967-4_7

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free