Joint Communication and Processor Frequency Selection for Low-Energy Systems under Timing Constraints  

作　　者：WANG Yingfeng LIU Zhijing 

机构地区：[1]School of Computer Science and Technology, Xidian University, xi＇an, China

出　　处：《China Communications》2010年第4期132-136,共5页中国通信（英文版）

摘　　要：This paper presents a method to reduce the energy consumption of multi-core systems characterized by processor cores and buses with discrete frequency levels under timing constraints.The proposed method takes the transformations of the original task graphs,which include dependent tasks located in different iterations,as inputs.The proposed method utilizes mapping selection as well as joint processor and communication frequency scaling to implement energy reduction.We conduct experiments on several random task graphs.Experimental results show that the proposed method can achieve substantial energy reduction compared with previous work under the same hard timing constraints.This paper presents a method to reduce the energy consumption of multi-core systems characterized by processor cores and buses with discrete frequency levels under timing constraints. The proposed method takes the transformations of the original task graphs, which include dependent tasks located in different iterations, as inputs. The proposed method utilizes mapping selection as well as joint processor and communication frequency scaling to implement energy reduction. We conduct experiments on several random task graphs. Experimental results show that the proposed method can achieve substantial energy reduction compared with previous work under the same hard timing constraints.

关 键 词：Green Computing Dynamic Voltage Scaling MULTI-CORE Adaptive Body Bias 

分 类 号：TN91[电子电信—通信与信息系统] TP332[电子电信—信息与通信工程]