Research of Self-Adapted Scheduling Policies for Supercomputer Computing Resource Based on Service Differentiation

Article Preview

Abstract:

The scheduling of supercomputer resource can greatly affect the computing speed of simulation and improve the efficiency of resource. This paper raise a design of resource scheduling architecture, which has centralized- control and distributed agent design. Aiming at the importance of tasks, a self-adapted scheduling policy way of computing resource is presented, which is used for service differentiation. Then key algorithms are described. Finally, simulation results prove the effect of our policy design.

You might also be interested in these eBooks

Info:

Periodical:

Pages:

735-739

Citation:

Online since:

December 2012

Export:

Price:

Permissions CCC:

Permissions PLS:

Сopyright:

© 2013 Trans Tech Publications Ltd. All Rights Reserved

Share:

Citation:

[1] J.Mark Bull, Chris Johnson. Data Distribution, Migration and Replication on a cc-NUMA Architecute[M]. University of Edinburgh, Scotland, UK.

Google Scholar

[2] Jun Zhang, Zhengbin Pang, Yi Zheng, et al. Design and Implementation of a Block Transfer Communication Mechanism in CC-NUMA Systems[J]. Journal of Computer Engineering & Science, 2008,vol1.

Google Scholar

[3] Ruilu Yi, Fei Hu, Shuiqin Xi, et al. Distributed Parallel Computing Based on LSF Cluster System[J]. Measurement & Control Technology, 2006,vol25(7).

Google Scholar

[4] Yutong Lu, Min Zhu, Zhiyu Shen. Research on Resource Management Structure of Large Scale Parallel Computing. The 7th National Parallel Computing Conference. Chengdu, 2003.

Google Scholar

[5] Ruisheng Miao, Xianming Ju, Jiasheng Wu. Missile Aerodynamics[M]. National Defense Industry Press,2006.01.

Google Scholar

[6] Li Wan. Solutions for Transient Dynamics CAE& Example Tutorials of MSC.Dytran[M]. Peking University Press, Nov, 2004.

Google Scholar

[7] Zepeng Wang. ANSYS 13.0/LS-DYNA Example Tutorials of Nonlinearized Finite Element Analysis[M]. China Machine Press, 2011.

Google Scholar