Abstract—A parallel computing system helps the user to exploit the parallelism inherent in the job to ensure its execution in the minimum time. This is done by distributing the job over multiple processors available in the system. Scheduling a job on such a system gains paramount importance and is always desired from a scheduling policy. Efficient job execution exploits the software parallelism in the job to map it with the available parallel hardware. A batch scheduler schedules the similar jobs by pooling them together and allocating them on the appropriate processor. Batch scheduling is useful as it allows the sharing of resources among many users and programs. Further, it avoids the idling of resources thereby increasing the utilization to a greater extent. A batch scheduler ensures proper amalgamation of software parallelism with hardware parallelism. This paper compares the performance of three batch scheduling policies viz. First Come First Serve (FCFS), Turnaround Based Scheduling Scheme (TBSS) and Batch Scheduling Scheme (BS). Simulation study is performed to analyze the performance of these three strategies under various test conditions that involves varying the hardware and software parallelism to observe the effect on the Turnaround Time (TAT) for the batch of jobs.
Index Terms—Batch scheduling, JPDG, parallel computing, turnaround time (TAT).
The authors are with the School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi - 110067, India (e-mail: mdshahid.cs@ gmail.com, firstname.lastname@example.org, email@example.com).
Cite: Mohammad Shahid, Zahid Raza, and Deo Prakash Vidyarthi, "A Comparative Study of Batch Scheduling Strategies for Parallel Computing System," International Journal of Innovation, Management and Technology vol. 4, no. 1, pp. 31-37, 2013.