An efficient and effective performance evaluation method for multiprogrammed multiprocessor systems
Year: 1997
Type: conference-paper
Abstract: A partitionable shared memo~, multiprocessor system can be partitioned into many virtual subsystems that can be allocated to independent jobs.A multi-server queueing system model for approximating performance of partitionable multiprocessor systems is proposed in this paper.A system is specified by the number of processors, and jobs are characterized by a probability distribution of the number of processors required, a Poisson arrival process, and an exponential distribution of execution time.The first-come-first-served (FCFS) queueing discipline is analyzed.An adapted FCFS policy is also considered, which, when there are no enough processors available for large jobs in the front of the waiting queue, schedules smaller jobs behind these large jobs earlier.Performance measures reported include mean job waiting and response times, average queue length, and system utilization.Numerical data for a system consisting of 32 processors are provided.Simulation results validating our model are also presented.
Author: Keqin LI
Institutions: State University of New York, SUNY New Paltz
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W2114938967
