基于混合优化的多核处理器动态热管理方法

Hybrid Dynamic Thermal Management Method for Multi-core Microprocessors

对于多核/众核处理器,芯片功耗增加导致温度急剧升高。同时,芯片热管理的控制复杂度随着处理器核心数量的增加呈指数增长关系,传统的热管理技术已经无法满足当前多核/众核处理器的应用需求。提出一种基于混合优化的多核处理器动态热管理方法,通过微分方程热建模、基于模型预测控制的功耗分析,以及混合优化的核内/核间热管理方法,实现多核/众核处理器芯片高效可靠运行。实验结果显示,芯片各核心温度均维持在安全温度附近,并且芯片核心间的温度差别较小,处于高效运行状态。

For multi-core/many-core microprocessors, the thermal increases rapidly as the power increases. Meanwhile, it is challenging to manage the thermal behavior of many-core microprocessors while still keeping them running at high performance since the control complexity increases as the core number increases. A hybrid dynamic thermal management method is proposed. The new method employs differential equation thermal model, model predictive control, and hybrid thermal management method. Experimental results show that, the proposed method can keep the chip in safe temperature range, and retain the system in high performance.