行并行可重构单元阵列流水映射性能评估  

Pipeline Mapping Performance Evaluation for Row Parallel Reconfigurable Cell Array

作　　者：陈乃金[1,2] 冯志勇 江建慧[3] 何瑞祥[2] 王真 CHEN Naijin FENG Zhiyong JIANG Jianhui HE Ruing WANG Zhen(School of Computer Science and Technology, Tianjin University, Tianjin 300072, China College of Computer and Information Science, Anhui Polytechnic University, Wuhu, Anhui 241000, China School of Software Engineering, Tongji University, Shanghai 201804, China School of Computer Science and Technology, Shanghai University of Electric Power, Shanghai 200090, China)

机构地区：[1]天津大学计算机科学与技术学院,天津300072 [2]安徽工程大学计算机与信息学院,安徽芜湖241000 [3]同济大学软件学院,上海201804 [4]上海电力学院计算机科学与技术学院,上海200090

出　　处：《同济大学学报（自然科学版）》2017年第8期1218-1226,共9页Journal of Tongji University:Natural Science

基　　金：国家"八六三"高技术研究发展计划(2013AA013204);国家自然科学基金(61432017;61572036);安徽省自然科学基金(1408085MF124);安徽省高校自然科学研究基金重点项目(KJ2015A003;KJ2013A001);安徽省高校优秀中青年骨干人才国内外访学研修重点项目(gxfxZD2016102)

摘　　要：针对粗粒度单元阵列流水映射问题,设计了三种行流水结构阵列,并分析了其执行步骤,提出了一种基于行流水阵列通用的流水映射算法.该算法综合考虑混合多层迭代启动间距、块间流水通信成本、块配置成本等多个因素,一组测试基准程序实验结果表明了文中算法的合理性,与多目标优化映射算法相比,该算法消耗总时延平均节省了4.0%(可重构单元阵列RCA_(4×4))和4.3%(可重构单元阵列RCA_(8×8));与满射映射相比,该算法消耗总时延平均节省了52.1%(RCA_(4×4))和56.2%(RCA_(8×8)).As for the problem of coarse-grained cell array pipeline mapping,this paper designed three row pipeline architecture array,analyzed their execution step,and presented a universal pipeline mapping（PM）algorithm for row pipeline array.This algorithm had comprehensive considered multilevel iteration initiation interval,communication costs between blocks,block reconfigurable costs and etc.The experimental results of a set of benchmark programs show the rationality of the algorithm.Comparing with multi-objective optimization map（MOM）,the average execution total cycles of PM saved by 4.0%（reconfigurable cell array,RCA4×4）and 4.3%（reconfigurable cell array, RCA8×8）. Comparing with epimorphism map（EPIMap）algorithm,the average execution total cycles of PM saved by 52.1%（RCA4×4）and 56.2%（RCA8×8）.

关 键 词：行流水 映射 多约束 流水段 启动间距 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]