期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2018
卷号：18
期号：12
页码：125-130
出版社：International Journal of Computer Science and Network Security
摘要：The energy and power of the multicore system is controlled and measured by the hardware management system. Power densities are proportionally increasing due to increase in temperature of the multi-processor system on chip MPSOc. Due to shrinking of the chip size and increased number of transistor on chip high peak temperature directly affect the speed and life span of the multicore system. Migration of task that produces high thermal value to such cores that has less temperature is a common technique to avoid high thermal values in multi core system. The proposed offline Scheduling Algorithm technique migrate task load to the core that has less thermal values in response to a threshold temperature this technique also considers other thermal problems which affects the power, reliability and performance of multi-core system. For evaluating and attaining the Stability of the proposed research for migration of tasks can be verified by comparing it with the different task migration techniques that are familiarized previously. The proposed technique also considers both hot and cold tasks to forecast the accurate thermal values by considering the previous thermal history of the tasks. The scheduling mechanism only considering such cores that are in highest energy state due to execution of some task such as running state while those cores that are in sleep or deep sleep mode can be consider in a lowest energy state. The simulation results shows that the proposed technique reduces almost 6℃ temperatures at 13% utilization and the energy utilization is 230.43 J which is 5.12 % improved as compare to the global EDF used previously. The proposed scheduling technique can control and measure the thermal effects and energy consumption on chip and utilize the processor in an efficient way when least number of tasks is in running state but has the ability to schedule multiple task to make a slower and energy efficient processor.
关键词：Dynamic Voltage and Frequency Scaling; Dynamic Power Management; Dynamic Thermal Management; Dynamic Voltage Scaling; Energy Optimization
