A Dynamic Scheduling Algorithm for Spawn Processes in MPI-2 to Improve and Maintain Load Balancing
|
10.5120/18569-6055 |
Sherif Abdelazim Embaby, Ahmed Shawky Moussa and Ibrahim Farag. Article: A Dynamic Scheduling Algorithm for Spawn Processes in MPI-2 to Improve and Maintain Load Balancing. International Journal of Computer Applications 106(12):1-6, November 2014. Full text available. BibTeX
@article{key:article,
author = {Sherif Abdelazim Embaby and Ahmed Shawky Moussa and Ibrahim Farag},
title = {Article: A Dynamic Scheduling Algorithm for Spawn Processes in MPI-2 to Improve and Maintain Load Balancing},
journal = {International Journal of Computer Applications},
year = {2014},
volume = {106},
number = {12},
pages = {1-6},
month = {November},
note = {Full text available}
}
Abstract
Some parallel applications that solve big problems in fields like weather forecasting, data analysis, energy fields, and protein folding need to create unpredicted processes at the application run time. The MPI package provides capability to not only write static parallel programs but also to create dynamic processes at run time. However, the MPI standard did not provide any way to schedule these dynamically created processes. Online scheduling can be a solution for this problem. Hence, this work introduces an online scheduling algorithm for dynamically created processes on cluster's nodes according to the machines performance. The objective of the algorithm is to achieve dynamic load balancing along the scheduler run time over heterogeneous cluster's machines. In addition, the proposed scheduler achieves load balancing over heterogeneous hardware according to the current real-time state of the nodes even if the processors are responding to other parallel simultaneous schedulers or running other parallel or sequential programs.
References
- "mpi-2: Extensions to the message passing interface" message passing interface forum, 1997. http://www. mpiforum. org.
- M´arcia C Cera, Guilherme P Pezzi, Elton N Mathias, Nicolas Maillard, and Philippe O a Navaux. Improving the dynamic creation of processes in MPI-2. Lecture Notes in Computer Science, 4192/2006:247–255, 2006.
- MPICH Organization. http://www. mpich. org/.
- Pavan Balaji, Darius Buntinas, David Goodell, William Gropp, Jayesh Krishna, Ewing Lusk, and Rajeev Thakur. PMI : A Scalable Parallel Process-Management Interface for Extreme-Scale Systems.
- William Gropp, David Ashton, Darius Buntinas, and Brian Toonen. MPICH2 Installers Guide. 2009.
- M´arcia C Cera, Guilherme P Pezzi, Nicolas B Maillard, and Philippe O A Navaux. Scheduling Dynamically Spawned Processes in MPI-2.
- Edgar Gabriel, Graham E Fagg, and Jack J Dongarra. Evaluating the Performance of MPI-2 Dynamic Communicators and One-Sided Communication. pages 88–97, 2003.
- Peter Brucker and P Brucker. Scheduling algorithms, volume 3. Springer, 2007.
- William Saphir, Leigh Ann Tanner, and Bernard Traversat. Job Management Requirements for NAS Parallel Systems and Clusters. NAS Technical Report NAS-95-006, pages 1–19, 1995.
- Subramanian Kannan, Mark Roberts, Peter Mayes, Dave Brelsford, and Joseph Skovira. Workload Management with LoadLeveler. IBM Corporation, International Technical Support Organization, 2001.
- Condor Team. Condor R version 6. 8. 2 manual. 2006.
- Songnian Zhou, Xiaohu Zheng, Jingwen Wang, and Pierre Delisle. Utopia: a load sharing facility for large, heterogeneous distributed computer systems. Software: Practice and Experience, 23(12):1305–1336, 1993.
- W Gropp and E Lusk. Dynamic process management in an MPI setting. ProceedingsSeventh IEEE Symposium on Parallel and Distributed Processing, pages 530–533, 1995.
- Message Passing and Interface Forum. MPI : A Message-Passing Interface Standard. Forum American Bar Association, 8(UT-CS-94-230):647, 2009.