Research On Task Scheduling Of Multi-core Processors Based On Dynamic Adaptive Strategy

As the mainstream of microprocessors’development, multi-core processors become widely popular, thanks to their high efficiency and low cost characteristics. The utilization of multi-core technology increases the performance of microprocessors significantly. In the course of the study of multi-core processor system, tasks scheduling is particularly important. Furthermore, it puts forward higher requirements on the scheduling strategy. At the same time, the scheduling problem of multi-core systems has become a hot issue in the field of IT technology, with the rapid development of multi-core systems. In this new era, the research focus is on how to improve the systems’efficiency and how to keep load balance while the tasks run in parallel.First, this thesis describes the background and the significance of the research, introduces briefly the basic theoretical knowledge on task scheduling of multi-core systems and modeling based on queuing theory. It is difficult to make a comprehensive evaluation on scheduling algorithm based on a single performance indicator, so we consider the average turnaround time, average response time and the utility of the processor in this paper. A new approach combining multi-objective decision theory is proposed and we have obtained the membership linear weighted average. It provides a theoretical basis for designing the optimal strategy selection module.And then, the thesis researches several typical tasks scheduling algorithms and analyzes the advantages and disadvantages of each algorithm. Besides, on the basis of the typical algorithms, it proposes dynamic adaptive scheduling algorithm. Based on this approach, a dynamic adaptive scheduling model is proposed and implemented, which takes advantage of existing scheduling algorithms to generate dynamic adaptive scheduling schemes and select the scheme that can provide optimal membership linear weighted average for applications. Experimental results show that the proposed model can provide dynamic adaptive scheduling and the adaptive service scheme is further superior in some occasion.Finally, the dynamic adaptive scheduling algorithm based on load balancing is put forward respectively combined of load balancing knowledge. The new algorithm can reduce the average response time of tasks and improve the utility of the processor evenly.For the purpose of verify the performance of improved algorithm, this paper implements this algorithm. Moreover, the corresponding experiments on the feasibility and comparison of similar algorithms are done under the environment of Microsoft Visual C++6.0. After that, this paper gives the result analysis and finally proves that the improved algorithm can effectively solve tasks scheduling problem for the multi-core processor system.