Job Scheduling Simulator for Assisting the Mapping Configuration Between Queue and Computing Nodes
In computer centers responsible for providing users with high-performance computing resources such as computer clusters, job throughput is an important criterion in satisfying users’ computing needs. The setting and configuration parameters in a job scheduler affect the criterion. In particular, the mapping between queues deployed in the job scheduler and computing nodes is highly related to job throughput. In most cases, however, the mapping configuration is conducted based on the administrators’ experience and knowhow partly because the tools that facilitate the determination of the mapping are not available. In this paper, we propose a job scheduling simulator that allows the administrators to investigate how the mapping affects job throughput. In the evaluation, the behavior of the proposed job scheduling simulator is assessed through a comparison with an actual computer cluster. In addition, real cases using the proposed job scheduling simulator are discussed.
This work was partially supported by JSPS KAKENHI Grant Numbers JP16H02802, JP17K00101.
- 3.Cybermedia Center, Osaka University: Large-Scale Computer System (2018). http://www.hpc.cmc.osaka-u.ac.jp/en
- 5.Gentzsch, W.: Sun grid engine: towards creating a compute power grid. In: Proceedings of the 1st IEEE/ACM International Symposium on Cluster Computing and the Grid, pp. 35–36. IEEE (2001)Google Scholar
- 6.Green, T.P., Snyder, J.: DQS, a distributed queuing system. Florida State University (1993)Google Scholar
- 7.Henderson, R.L.: Job scheduling under the portable batch system. In: Workshop on Job Scheduling Strategies for Parallel Processing, pp. 279–294. Springer (1995)Google Scholar
- 8.Klusáček, D., Rudová, H.: Alea 2: job scheduling simulator. In: Proceedings of the 3rd International ICST Conference on Simulation Tools and Techniques, pp. 1–10. ICST (2010)Google Scholar
- 9.Obaida, M.A., Liu, J.: Simulation of HPC job scheduling and large-scale parallel workloads. In: Proceedings of the 50th Winter Simulation Conference (WSC), pp. 920–931. IEEE (2017)Google Scholar
- 11.Simakov, N.A., Innus, M.D., Jones, M.D., DeLeon, R.L., White, J.P., Gallo, S.M., Patra, A.K., Furlani, T.R.: A slurm simulator: implementation and parametric analysis. In: Proceedings of the 8th International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, pp. 197–217. Springer (2017)Google Scholar
- 12.TOP500.org: TOP500 The List (2018). http://www.top500.org
- 14.Tulasi, B., Wagh, R.S., Balaji, S.: High performance computing and big data analytics - paradigms and challenges. Int. J. Comput. Appl. 116(2), 28–33 (2015)Google Scholar
- 15.Yoo, A.B., Jette, M.A., Grondona, M.: SLURM: simple linux utility for resource management. In: Proceedings of the 9th Workshop on Job Scheduling Strategies for Parallel Processing, pp. 44–60. Springer (2003)Google Scholar