Alen Grižonič (2009) The Simulation of Process Scheduling Algorithms. EngD thesis.
The use of computer systems is gradually entering lives of numerous people. The quicker as the speed of life is, faster and more efcient are the devices surrounding us. The need for quick communication and the execution of demanded tasks confront us with a fact that the use of the mentioned means is almost inevitable and completely sensible. The question, however, remains what gives an assurance of optimal realization. Computers are so widely used that several developers offer more and more effcient operating systems, which enable the execution of user requests, and act as user interfaces between the individual and computer hardware at the same time. Whether algorithms used by different operating systems currently present on the market are comparable, and/or their speed of process scheduling can be measured and adapted to a great deal of different system demands, is questionable. For this purpose I carried out a simulation showing how two different up to now competitive operating systems use their computer system resources based on a multiprocessor computer architecture. The first part of the thesis presents the processes inside the system, the effciency evaluation criteria used, and the basic principals of process scheduling. In the second part I then concentrate on the most commonly used algorithms, the notion of real-time scheduling, and the advantages of the multiprocessor or multicore systems respectively. What follows is a detailed description of the algorithms for process scheduling and the most important, comparison and implementation of the simulation, which (with the use of the results acquired) helped me establish with which system load the particular system or its algorithm proves the most effective.
Actions (login required)