A Fast-Start, Fault-Tolerant MPI Launcher on Dawning Supercomputers

Author

Liu, Xu ; Tu, Bibo ; Zhan, Jianfeng ; Meng, Dan

Author_Institution

Nat. Res. Center for Intell. Comput. Syst., Chinese Acad. of Sci., Beijing

fYear

2008

fDate

1-4 Dec. 2008

Firstpage

263

Lastpage

266

Abstract

Daemon-based MPI launchers are the mainstream in nowadays, because they can startup processes rapidly. However, effective task management and fault tolerance become more important as the scale of supercomputers enlarges. A new fast-start and fault tolerant launcher, called SFLauncher, has been used to startup MPICH task on Dawning supercomputers. This paper details its features and implementation, with emphasis on scalability, self-organization algorithm and garbage reclamation. The results of performance evaluation on SFLauncher are also given.

Keywords

application program interfaces; fault tolerance; message passing; parallel machines; dawning supercomputer; fast-start fault-tolerant MPI launcher; message passing interface; task management; Computer crashes; Distributed computing; Fault tolerance; Fault tolerant systems; Heart beat; Intelligent systems; Partitioning algorithms; Peer to peer computing; Scalability; Supercomputers; MPI launcher; fault tolerance; scalability;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Computing, Applications and Technologies, 2008. PDCAT 2008. Ninth International Conference on

Conference_Location

Otago

Print_ISBN

978-0-7695-3443-5

Type

conf

DOI

10.1109/PDCAT.2008.56

Filename

4710990