S124 The Effectiveness of a Raspberry Pi Cluster to Perform WRF Simulations for Enhancing Computational Resources within the Valparaiso University Meteorology Program

Sunday, 6 January 2019
Hall 4 (Phoenix Convention Center - West and North Buildings)
Maxwell Grover, Valparaiso University, Valparaiso, IN; and I. Arseneau, E. Foust, T. Wade, and K. H. Goebbert

This project assesses the performance characteristics of an inexpensive computer cluster and a standard multi-core computer server that are capable of running the Weather Research and Forecasting (WRF) Model to study various weather phenomenon from winter storms to landfalling hurricanes and severe thunderstorm outbreaks. Multiple Raspberry Pis, small computers that are relatively inexpensive, are connected to create an inexpensive compute cluster which could have 40 to 50 cores, that will allow for substantial weather simulations to be completed. The Raspberry Pi cluster performance will be compared to a 15-core server to assess its effectiveness in running the WRF. Specifically, the comparison of of clock-time completion of numerical simulations will be assessed. This paper will describe the viability of using the relatively inexpensive cluster for use in teaching and research within the Valparaiso University meteorology program. This work was inspired by similar work being done at the National Center for Atmospheric Research.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner