WRF, análisis de rendimiento en clústeres HPC
Keywords:
WRF, HPC, performance, Infiniband, parallel, scalability, CCAbstract
The WRF (Weather Research and Forecasting) model is a numerical weather prediction system, designed for both forecasting and atmospheric research, with an emphasis on scalability, efficiency and portability. It has been deployed successfully in HPC clusters. Therefore, the understanding of the dependency of the WRF with the different hardware elements is crucial to make efficient predictions. Because of this we analyzed the scalability of the WRF based on three parameters, respectively MPI implementations, inter-node communication speeds, and MPI processes per processor, enabling benchmarking that allows to understand the relationship between scalability and these three parameters. The results show a dependence of the scalability of the WRF with the inter-node communications; consequently, when using high-speed networks, such as Infiniband, a scalability higher than the 2 computational nodes was obtained, which was the maximum scalability achieved when using Networks with lower speeds like Ethernet.
Downloads
Metrics
References
Atchley, S., Dillow, D., Shipman, G., Geoffray, P., Squyres, J. M., Bosilca, G., Minnich, R. (2011). The Common Communication Interface (CCI). En: IEEE 19th Annual Symposium on High Performance Interconnects, pp. 51-60. https://doi.org/10.1109/HOTI.2011.17
Gualán, R., Solano-Quinde, L. (2014). Análisis de rendimiento y profiling del modelo WRF en un clúster HPC. En: Actas del Congreso TIC.EC 2014, pp. 151-162. https://publicaciones.ucuenca.edu.ec/ojs/index.php/maskana/article/view/730.
HPC Advisory Council. (2017). HPC Advisory Council - Best Practices. Retrieved September 22, 2017 from http://hpcadvisorycouncil.com/best_practices.php
Intel ®. (n.d.). Intel® Trace Analyzer and Collector | Intel® Software. Retrieved June 16, 2017 from https://software.intel.com/en-us/intel-trace-analyzer
Intel ®. (2017). Intel® MPI Library | Intel® Software. Retrieved June 18, 2017 from https://software.intel.com/en-us/intel-mpi-library
Michalakes, J., Loft, R., Bourgeois, A. (2001). Performance-portability and the Weather Research and Forecast Model. 11 p. Disponible en http://citeseerx.ist.psu.edu/viewdoc/ download?doi=10.1.1.8.3781&rep=rep1&type=pdf
OpenFabrics Alliance. (2015). OpenFabrics Software. Retrieved May 14, 2017 from https://www.openfabrics.org/index.php/openfabrics-software.html
Shainer, G., Liu, T., Michalakes, J., Liberman, J., Layton, J., Celebioglu, O., … Cownie, D. (2009). Weather Research and Forecast (WRF) Model: Performance Analysis on Advanced Multi-core HPC Clusters. The 10th LCI International Conference on High Performance Clustered Computing, 1-14.
Skamarock, W. C., Klemp, J. B., Dudhi, J., Gill, D. O., Barker, D. M., Duda, M. G., … Powers, J. G. (2008). A description of the advanced research WRF Version 3. NCAR Tech, 113. https://doi.org/10.5065/D68S4MVH
The Open MPI Project. (2016a). FAQ: Running MPI jobs. Recuperado el 10 de mayo de 2017 a partir de https://www.open-mpi.org//faq/?category=running#oversubscribing
The Open MPI Project. (2016b). FAQ: Tuning the run-time characteristics of MPI OpenFabrics communications (InfiniBand, RoCE and iWARP). Recuperado el 18 de junio de 2017 a partir de https://www.open-mpi.org/faq/?category=openfabrics#ib-components
Downloads
Published
How to Cite
Issue
Section
License
Copyright © Autors. Creative Commons Attribution 4.0 License. for any article submitted from 6 June 2017 onwards. For manuscripts submitted before, the CC BY 3.0 License was used.
You are free to:
 |
Share — copy and redistribute the material in any medium or format |
 |
Adapt — remix, transform, and build upon the material for any purpose, even commercially. |
Under the following conditions:
 |
Attribution — You must give appropriate credit, provide a link to the licence, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licenser endorses you or your use. |
| No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the licence permits. |
