[1]陈红梅,郭 伟,赖重远.小型通用高性能计算平台的设计与实现[J].计算机技术与发展,2019,29(10):186-190.[doi:10. 3969 / j. issn. 1673-629X. 2019. 10. 036]
　CHEN Hong-mei,GUO Wei,LAI Zhong-yuan.Design and Implementation of Small Universal High Performance Computing Platform[J].,2019,29(10):186-190.[doi:10. 3969 / j. issn. 1673-629X. 2019. 10. 036]
点击复制

小型通用高性能计算平台的设计与实现()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
29
期数:
2019年10期
页码:
186-190
栏目:
智能、算法、系统工程
出版日期:
2019-10-10

文章信息/Info

Title:
Design and Implementation of Small Universal High Performance Computing Platform
文章编号:
1673-629X(2019)10-0186-05
作者:
陈红梅1 ; 郭 伟2 ; 赖重远1
1. 江汉大学 交叉学科研究院,湖北 武汉 430056; 2. 江汉大学 数学与计算机科学学院,湖北 武汉 430056
Author(s):
CHEN Hong-mei 1 ; GUO Wei 2 ; LAI Zhong-yuan 1
1. Institute of Interdisciplinary Research,Jianghan University,Wuhan 430056,China; 2. School of Mathematics and Computer Science,Jianghan University,Wuhan 430056,China
关键词:
大数据处理; 高性能计算平台; 多核 CPU; GPU; SMP
Keywords:
big data processing; high performance computing platform; multi-core CPU; GPU; SMP
分类号:
TP399
DOI:
10. 3969 / j. issn. 1673-629X. 2019. 10. 036
摘要:
随着高性能计算技术和应用的发展,计算模拟已成为科学研究中不可缺少的第三种方法。 国内各大高校和科研院所纷纷建立了高性能计算平台,它们大多是针对重点学科大规模数据处理的特点,建立了比较有针对性的高性能计算平台。 江汉大学作为一所综合性普通高等学校,其大数据处理的规模相对有限、类型多样,且时间具有阶段性。 依据自身大数据处理的特点,本着开放共享、全体受益的原则,江汉大学设计并实现了小型通用高性能计算平台的建设。 该平台创建了一个通用的计算环境,同时又考虑了不同用户和应用的特殊需求。 平台提供多核 CPU 并行计算、GPU 并行计算和SMP 并行计算。 平台建立后为各院系提供了高性能计算服务,支持了学校多个科研项目,大力支援了学校科研工作的发展。
Abstract:
With the development of high performance computing technology and application,computational simulation has become an indispensable third method in scientific research. Domestic universities and research institutes have established high performance computing platform,most of which are aimed at big data processing of key disciplines and have established a relatively high performance computing platform. As a comprehensive university,the big data processing in Jianghan University is relatively limited, various types and its time is intermittent. According to the characteristics of its own big data processing,in line with the principle of open sharing and benefit for all, Jianghan University has designed and implemented a small universal high performance computing platform. This platform creates a universal computing environment while taking into account the special needs of different users and applications. The platform provides multi-core CPU parallel computing,GPU parallel computing and SMP parallel computing. After the establishment of the platform,it provides high performance computing services for each department,supports many scientific research projects of the university,and strongly supports the development of scientific research work of the university.

相似文献/References:

[1]李聪颖[],王瑞刚[],梁小江[]. 基于Hadoop的交互式大数据分析查询处理方法[J].计算机技术与发展,2016,26(08):134.
　LI Cong-ying[],WANG Rui-gang[],LIANG Xiao-jiang[]. An Interactive Processing Method of Analysis and Query for Big Data Based on Hadoop[J].,2016,26(10):134.
[2]陈园园,崔贯勋.高性能计算平台建设、运行与服务模式的探索[J].计算机技术与发展,2021,31(07):87.[doi:10. 3969 / j. issn. 1673-629X. 2021. 07. 015]
　CHEN Yuan-yuan,CUI Guan-xun.Exploration of Construction,Operation and Service Mode of HPC Platform[J].,2021,31(10):87.[doi:10. 3969 / j. issn. 1673-629X. 2021. 07. 015]

更新日期/Last Update: 2019-10-10