基于线程与分布式排序对比实验的设计与研究  

Design and research of comparing experiment based on thread and distributed sorting

作　　者：张国华[1] 叶苗[1] 陆霞[1] 吉晓香 梁德鸿 ZHANG Guohua;YE Miao;LU Xia;JI Xiaoxiang;LIANG Dehong(Taizhou College,Nanjing Normal University,Taizhou 225300,China)

机构地区：[1]南京师范大学泰州学院,江苏泰州225300

出　　处：《实验技术与管理》2020年第8期186-188,199,共4页Experimental Technology and Management

基　　金：国家自然科学基金青年基金项目(51708265);江苏省高校自然科学研究面上项目(19KJD520008);江苏省高等学校大学生创新创业训练计划项目“基于零销售业销售大数据分析与研究”;泰州市科技支撑计划项目(TS201927)。

摘　　要：为解决传统单机排序算法无法高效排序海量数据问题,帮助学生掌握单机单线程、多线程及分布式编程的区别及特点,并能灵活选择相应技术,文章设计了单线程、多线程及分布式环境下排序对比实验。该实验对这三种编程法的执行流程、算法代价、适用场景等进行了分析研究,并根据实验结果得出:在分布式环境下,为节约集群规模,可在计算节点上采用多线程方法来有效解决海量数据的处理问题。In order to solve the problem that traditional single-machine sorting algorithm can’t sort massive data efficiently,to help students understand the differences and characteristics of single-machine single-thread,multi-thread and distributed computing,and to choose suitable scenarios,this paper designs a single-thread,multi-thread and distributed environment sorting comparing experiment,and executes these three experiments.Processes,algorithm costs and applicable scenarios are analyzed and studied.According to the experimental results,in the distributed environment,and in order to save the cluster size,multithread method can be used on the computing nodes to effectively solve the problem of massive data processing.

关 键 词：排序 多线程 分布式计算 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]