Peak power reduction methodology for multi-core systems

In this paper, we propose the task scheduling for preventing the occurrence of peak power in the multi-core systems considering data dependence. In the mobile system, the peak power reduces the battery lifetime and makes the system unstable. Among the power consumption of system, the proportion of power consumption of core is very large. If cores execute multiple tasks simultaneously, this gives rise to the peak power. Therefore, algorithm to minimize the occurrence of peak power is needed. When multiple tasks are allocated to the multi-core systems, data dependence relations of all tasks should be considered to avoid data interferences. The proposed algorithm to reduce the peak power schedules the tasks with the data dependence information after data dependence analysis. The proposed algorithm is composed of task partitioning step, data dependence analysis step and priority scheduling step. The simulation results show that the proposed algorithm reduce the occurrence of peak power by up to around 11% compared to the existing algorithms.