Simulation of Multiprocessor System Scheduling

The speed and performance of computers have become a major concern today. Multiprocessor systems are introduced to share the work load between the processors. By sharing the work load among processors, the speed and efficiency of the system can be increased drastically. To share the workload properly between the processors in the multiprocessor system a proper scheduling algorithm is needed. Hence, one of the major factors that influence the speed and efficiency of the multiprocessor system is scheduling. In this thesis, the main focus is on the process scheduling for multiprocessor systems. The factors which influence scheduling and scheduling algorithms are discussed. Based on this idea of sharing the load among processors in the multiprocessor system, a simulation model for scheduling in a symmetric multiprocessor system is developed in the Department of Digital and Computer systems at Tampere University of Technology. This model treats all the processors in the system equally by allocating equal processor time for all the processes in the task and also evaluates the total execution time of the system for processing an input job. The scheduling algorithm in this simulation model is based on the input processes priority. The necessity of scheduling in multiprocessor systems is elaborated. The goal of this thesis is to analyse how process scheduling influences the speed of the multiprocessor system. Also, the difference in total execution time of the input jobs with different number of processors and capacity of the processors in the multiprocessor system is studied