Algorithmic approach for applying load balancing during task migration in multi-core system

As day by day multi core systems are increasing researchers are looking for various techniques to reduce power consumption and improve performance. In multi-core processors several cores are placed on the single grid and multiple tasks are distributed over these cores for processing [1]. Each core is responsible for the processing of the task allocated to it. This may lead to problems like load balancing on different cores, as various tasks are allocated to the single core. Now if there is a situation that a single core or a cluster of cores have large number of task to be executed while some of the cores remain free then there is a need of a technique called task migration. Task migration is not similar to load balancing in any aspect. In task migration only the migration of the task from overloaded core to other core is done but till now maintaining load balancing during task migration was an issue[2]. But if the load balancing is done at the time of the task migration then it will improve performance about 99%. This paper emphasis on various algorithms for load balancing during task migration in a multi-core system, so that number of tasks on different cores are distributed fairly. This paper also discusses the algorithms for deciding which task is to be migrated. With the help of these algorithms, performance and response time is improved to a great extent.