[1]黄锦增 陈虎 赖路双.异构GPU集群的任务调度方法研究及实现[J].计算机技术与发展,2012,(05):32-36.
 HUANG Jin-zeng,CHEN Hu,LAI Lu-shuang.Research and Implementation of Task Schedule Method on Heterogeneous GPU Cluster[J].,2012,(05):32-36.
点击复制

异构GPU集群的任务调度方法研究及实现()
分享到:

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

卷:
期数:
2012年05期
页码:
32-36
栏目:
智能、算法、系统工程
出版日期:
1900-01-01

文章信息/Info

Title:
Research and Implementation of Task Schedule Method on Heterogeneous GPU Cluster
文章编号:
1673-629X(2012)05-0032-05
作者:
黄锦增 陈虎 赖路双
华南理工大学软件学院
Author(s):
HUANG Jin-zeng CHEN Hu LAI Lu-shuang
School of Software Engineering, South China University of Technology
关键词:
负载均衡异构GPU集群任务调度动态适应
Keywords:
load balance heterogeneous GPU cluster task schedule dynamical adaptation
分类号:
TP311
文献标志码:
A
摘要:
GPU集群已经成为高性能计算的重要方式,特别对于计算密集型应用,具有成本低、性能高、功耗小的优势。为了解决GPU集群系统运行中的任务负载均衡问题,文中提出了一种面向计算密集型应用的异构GPU集群调度方法,该方法可以自动发现计算节点,并动态估计计算节点的计算能力,并根据计算能力、任务的计算强度和优先级在异构GPU集群上合理分配计算资源。同时,该系统还具有容错能力,能够处理计算节点的意外退出,可恢复意外退出计算节点的计算任务,并动态适应系统的计算规模。通过实验表明,文中采用的策略达到了预期目的
Abstract:
GPU cluster has become an important method for high performance computing, especially for compute-intensive applications. It has many advantages, such as low cost, high performance and low power consumption. To solve the load balancing problem of GPU cluster system, propose an algorithm for heterogeneous GPU cluster, it can automatically identify computation nodes, dynamically estimate the computing capability of these nodes and allocate resources in heterogeneous GPU cluster based on computation nodes" capability, tasks , computing strength and priority. At the same time, the system is also fault tolerant, which is able to handle unexpected exit of computa- tion nodes, recover the computing task of calculation nodes out of an unexpected exit and dynamically adapt to the calculation size of the system. The experiment result shows this strategy achieves desired purpose

相似文献/References:

[1]黄益贵 王汝传.P2P-VoD系统中自适应大小的滑动窗口模型研究[J].计算机技术与发展,2010,(05):21.
 HUANG Yi-gui,WANG Ru-chuan.Research of Self-Adjust Size of Sliding Window Model in P2P-Based VoD System[J].,2010,(05):21.
[2]徐群 祝永志.集群系统中的负载均衡问题的研究[J].计算机技术与发展,2009,(08):129.
 XU Qun,ZHU Yong-zhi.Research on Load Balancing Strategy for Cluster Systems[J].,2009,(05):129.
[3]陆磊 王锋.基于流负载均衡的入侵检测系统[J].计算机技术与发展,2009,(11):135.
 LU Lei,WANG Feng.Intrusion Detection System Based on Flow Load Balance[J].,2009,(05):135.
[4]王春枝 王骞.一种基于移动代理的服务器集群系统模型[J].计算机技术与发展,2009,(11):159.
 WANG Chun-zhi,WANG Qian.A Model of Servers Cluster System Based on Mobile Agent[J].,2009,(05):159.
[5]王洪臣 朱尚明.多连接校园网策略路由的研究与实现[J].计算机技术与发展,2008,(04):25.
 WANG Hong-chen,ZHU Shang-ming.Research and Implementation on Policy- Based Routing in Multihoming Campus Network[J].,2008,(05):25.
[6]刘必雄 许榕生[].大规模文件上传接收服务的负载均衡引擎研究[J].计算机技术与发展,2008,(06):16.
 LIU Bi-xiong,XU Rong-sheng.Research of Load - Balancing Engine for Large - Scale Up - Transfer Files Receiving Service[J].,2008,(05):16.
[7]李丙锋 祝永志 魏榕晖.异构Beowulf系统负载均衡技术的研究与实现[J].计算机技术与发展,2008,(07):60.
 LI Bing-feng,ZHU Yong-zhi,WEI Rong-hui.Implementation of Load Balancing Technology on Heterogeneous Beowulf System[J].,2008,(05):60.
[8]郭静 祝永志 王延玲.基于RR—DNS的Web集群系统的可用性研究[J].计算机技术与发展,2008,(12):56.
 GUO Jing,ZHU Yong-zhi,WANG Yan-ling.On Availability of Web Cluster Based on RR - DNS[J].,2008,(05):56.
[9]赵文评 葛玮.基于PDG图的分布式动态可执行服务组合方法[J].计算机技术与发展,2007,(07):40.
 ZHAO Wen-ping,GE Wei.Approach for Decentralizing Dynamic Web Services Composition Based on PDG[J].,2007,(05):40.
[10]阎文博 张育平 郭朝霞.基于网格技术的资源发现机制的研究与优化[J].计算机技术与发展,2007,(07):91.
 YAN Wen-bo,ZHANG Yu-ping,GUO Zhao-xia.Research and Amelioration on Resource Discovery Based on Grid Computing[J].,2007,(05):91.

备注/Memo

备注/Memo:
黄锦增(1987-),男,广东人,硕士研究生,研究方向为高性能计算;陈虎,副教授,博士,研究方向为计算机系统结构
更新日期/Last Update: 1900-01-01