Research On Multi-core Processor Temperature Equilibrium Strategy

With the progress of the times, subjected to the technology, the single-core processor performance is difficult to get promoted, and it is difficult to meet user’s demand for high-performance processors in life, Therefore, the concept of multi-core processors have been proposed, and now with the rapid development of multi-core processor, eight-core processors also have been widely used in daily life. However, with the continuous improvement of technology, more and more cores are integrated on the smaller and smaller chip, which would make the temperature of the processor generates more heat and bring the cooling problem.At the first of the paper, the present researches about temperature and power consumption are presented on different classification criteria. Based on the detailed analysis of them, mainly from the operating system point of view, especially from the application of Linux system, which is very widely to study the temperature and power equilibrium model on multi-core processors the main research content is based on the hardware performance counter to further calculate the process of power consumption predictive value. When the temperature of a processor’s inner core is higher than a threshold value, the weight of the process on the core of the processor is adjusted to shorten the running time of the hottest process in a time slice. If it fail to reach the effect of temperature control, it would based on the core of the processor load situation to nuclear transfer, so that the temperature distribution is more balanced, but also makes the temperature of the growth rate slowed down. The research goal of this model is in the premise of not reducing the processor performance, through the improvement of operating system scheduling strategy, make the temperature of processors distribution more uniform, and make the power consumption growth is more gentle.Finally, the paper through the HotSpot simulation tool to verify the research theory, which includes the core temperature control strategy and the effect of multi core process migration when the single core temperature is too high. Verification results show that this scheduling policy for improving, the temperature distribution of the equalizer processor has some effect, the system has some improved algorithms. The verification results show that the scheduling strategy has a certain effect on improving high-temperature core temperature and balancing the temperature distribution of the processor, and also improved the system algorithm.