Submitted as: model evaluation paper |
| 15 Dec 2017
Status: this preprint was under review for the journal GMD but the revision was not accepted.
Performance evaluation of ROMS v3.6 on a commercial cloud system
Kwangwoog Jung,Yang-Ki Cho,and Yong-Jin Tak
Abstract. Many commercial cloud computing companies provide technologies such as high-performance instances, enhanced networking and remote direct memory access to aid in High Performance Computing (HPC). These new features enable us to explore the feasibility of ocean modelling in commercial cloud computing. Many scientists and engineers expect that cloud computing will become mainstream in the near future. Thus, evaluation of the exact performance and features of commercial cloud services for numerical modelling is appropriate. In this study, the performance of the Regional Ocean Modelling System (ROMS) and the High Performance Linpack (HPL) benchmarking software package was evaluated on Amazon Web Services (AWS) for various configurations. Through comparison of actual performance data and configuration settings obtained from AWS and laboratory HPC, we conclude that cloud computing is a powerful Information Technology (IT) infrastructure for running and operating numerical ocean modelling with minimal effort. Thus, cloud computing can be a useful tool for ocean scientists that have no available computing resource.
Received: 29 Oct 2017 – Discussion started: 15 Dec 2017
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
The performance of the ROMS was evaluated on Amazon Web Services for various configurations. Our study shows how numerical ocean models can be constructed and parallelised in a commercial cloud computing environment and outlines how performance similar to local high-performance computing can be achieved in commercial cloud computing environments by optimising the modelling environment. Cloud computing can be a useful tool for those who have no available computing resource.
The performance of the ROMS was evaluated on Amazon Web Services for various configurations. Our...
- Printer-friendly version
- Supplement