PSO基础的链路平衡NoC映射与路由  

Mapping and routing of link-balance NoC based on PSO

作　　者：周文彪[1,2] 张岩[1] 毛志刚[2] 

机构地区：[1]哈尔滨工业大学深圳研究生院,广东深圳518055 [2]哈尔滨工业大学微电子中心,黑龙江哈尔滨150001

出　　处：《系统工程与电子技术》2008年第12期2506-2511,共6页Systems Engineering and Electronics

摘　　要：针对NoC设计流程中IP核映射和路由分配两个关键步骤,提出了一种基于PSO低功耗IP核映射和链路平衡的路由分配算法PLBMR。PLBMR是一个两步PSO优化算法,在建立NoC的通信功耗和链路负载的数学表示的基础上,采用一种有效IP核映射与路由分配的PSO粒子编码方法,使得NoC的通信能量消耗最小以及保证NoC的通信链路负载平衡。仿真结果表明,采用PLBMR算法产生的NoC通信能量消耗平均比GA和BnB分别少了0.7%和0.6%,特别是,PLBMR算法在计算时间上都优于GA和BnB;对于链路负载的平衡能力,PLBMR平均分别比MOCA小了9%,比BnB小了7%。Aiming at the two key steps of IP core, mapping and routing, in the NoC design flow, the paper presents a novel mapping and routing technique for the 2D mesh based NoC design problem with an objective of minimizing the energy consumption and normalized worst link-load. The proposed algorithm is a particle swarm optimization （PSO） based two phases process, one is mapping core onto NoC, and another is the allocation of the routing path. The proposed algorithm uses a novel representation for PSO particle which minimizes the NoC communication energy consumption and keeps the NoC link load balance. Experimental results show that PLBMR is 0.7% and 0.6% less than GA and BnB algorithm respectively on the NoC energy consumption, especially, the computation time of PLBMR is smaller than that of GA and BnB. And for the link-load balance, it is also 9% and 7 % better than MOCA and BnB respectively.

关 键 词：NOC 映射 路由 低功耗 链路负载平衡 

分 类 号：TN47[电子电信—微电子学与固体电子学]