计算机工程2017，Vol.43Issue(6)：1-10,18,11.DOI:10.3969/j.issn.1000-3428.2017.06.001
面向MPI集合操作的定制化片上网络
Customized Network-on-Chip Oriented to MPI Collective Operations
摘要
Abstract
According to the principle of computations approaching data,this paper proposes a design method of Networkon-Chip(NoC) oriented to MPI collective operations,which focuses on the hardware enhancement of common NoC routers to speed up MPI collective operations on the network layer.It designs MPI reduction,extends it to support more operations and combines it with an adaptive method for the deterministic routing algorithm,which can learn transmission paths of messages dynamically.Thus,enhanced routers can complete message processing in place,which not only speed up the processing procedure but also coalesce messages.The design method for detailed micro-architecture of NoC is presented.Different layout strategies of enhanced routers are compared and the corresponding performance,power consumption and extra chip-area are evaluated.Testing results show that,compared with ideal software-based implementation,the proposed method can improve the reduction performance by 6.4 ～ 41.7 times,broadcast by 15.3 ～ 31.2,global reduction by 5.4 ～ 9.7 times,and gather by 1.3 ～ 1.8 times,while the consumption of power and chip-area is limited.关键词
片上网络/片上多核处理器/消息传递接口/集合操作/定制化Key words
Network-on-Chip (NoC)/Chip Multi-Processor (CMP)/Message Passing Interface (MPI)/collective operation/customization分类
信息技术与安全科学引用本文复制引用
陆思羽,王宏伟,张悠慧,杨广文,郑纬民..面向MPI集合操作的定制化片上网络[J].计算机工程,2017,43(6):1-10,18,11.基金项目
国家“863”计划项目(2013AA01A215). （2013AA01A215）
