Nikitina T. Real-time scheduling methods for multicore processors

The object - the planning process and resource allocation algorithms in computer systems based on multicore processors. The subject - models and methods scheduling computing tasks for real-time multicore processors. The goal - the development of computational methods for planning real-time tasks that allow for shared resources multicore processor, which will improve the performance of real-time systems based on multicore processors. The model of real time system based on multi-core processors was improved which in contrast to the known takes into account shared cache multicore memory, multithreading structure of real-time software that allows to solve the problem of scheduling with proposed criteria of optimization.The first time the static scheduling method of real-time computational tasks, which takes into account multiple cores, shared cache multicore memory and by reducing the number of cache misses allows to increase real-time system performance based on multicore processors. The first time the dynamic scheduling method of real-time computational tasks, which takes into account multithreading periodic tasks with real-time constraints, multiple cores, shared cache multicore memory which allows to increase real-time system performance based on multicore processors. The method of exceptions scheduling problems for real-time systems based on multicore processors, which in contrast to analogue using a method based on approximate solution of the problem of the least coverage, so you can reduce the time complexity of the service queue of tasks. The exclusion scheduling method for real-time systems based on multicore processors was improved which in contrast to analogue using a method of solving the problem of minimum coverage, so it reduce the time complexity of the service queue of tasks.