Performance Efficient Heterogeneous Multi Core Scheduling Strategy Based on Genetic Algorithm

Multi-core processors offer a significant performance increase over single core processors. Therefore, they have the potential to enable computation-intensive real-time applications with stringent timing constraints that cannot be met on traditional single-core processors. However, with the number of cores on a single chip continuing to increase, it has been a great challenge to effectively manage the energy efficiency of multicore based systems. Power and temperature management are also two concerns that can increase exponentially with the addition of multiple cores. Design innovations in multicore processor architectures bring new optimization opportunities and challenges in the computing era. System performance will be further enhanced by addressing these challenges. In particular, the process (task) scheduler is one of the critical challenge is garnering great interest. High performance in a heterogeneous multicore system is essential which is achieved by effective scheduling, which remains a challenging problem. Further multi core technology opens research opportunities for energy reduction through efficient scheduling. There may be different hardware and software solutions for the above issue; hardware solutions are based on adjusting dynamic voltage per core, alternatively software approach includes, scheduling task among cores, in heterogeneous environment. Task scheduling in multicore architecture is an extremely difficult problem, because it requires a large combinatorial search space and also precedence constraints between the processes; for the effective utilization of multi core processor system, efficient assignment and scheduling of jobs is more important. Many of the existing algorithms are not focused on task scheduling and core utilization in heterogeneous multi core systems. This paper formulates task scheduling as an optimization problem and the results are compared with the earlier faster scheduler in use. Findings show that, in addition to its optimum solution for large scale problem, the Genetic Algorithm (GA) proposed here fits the heterogeneous multi core parallel scheduling problem of minimizing the completion time as well as in effective core utilization.