Research Article
Simulations on the Energy Consumption of WRF on Meteorological Cloud
@INPROCEEDINGS{10.1007/978-3-030-48513-9_47, author={Junwen Lu and Yongsheng Hao and Xianmei Hua}, title={Simulations on the Energy Consumption of WRF on Meteorological Cloud}, proceedings={Cloud Computing, Smart Grid and Innovative Frontiers in Telecommunications. 9th EAI International Conference, CloudComp 2019, and 4th EAI International Conference, SmartGIFT 2019, Beijing, China, December 4-5, 2019, and December 21-22, 2019}, proceedings_a={CLOUDCOMP}, year={2020}, month={6}, keywords={Moldable parallel tasks Energy consumption WRF Simulations Energy efficiency}, doi={10.1007/978-3-030-48513-9_47} }- Junwen Lu
Yongsheng Hao
Xianmei Hua
Year: 2020
Simulations on the Energy Consumption of WRF on Meteorological Cloud
CLOUDCOMP
Springer
DOI: 10.1007/978-3-030-48513-9_47
Abstract
In the paper, we try to evaluate the energy consumption of meteorological applications on meteorological cloud of on different kinds of processors. We take WRF (Weather Research and Forecasting model) model as the typical model. Three major factors are including in the evaluation: the energy consumption, the execution time, and the parallelism. The moldable parallel tasks have a scope of parallelisms. But after the job has an execution state, and the parallelism cannot be changed during the execution. Different to most of past research, our system support slots time and every job needs a few slot times to execute it. We give a detailed analysis of DVFS (Dynamic Voltage and Frequency Scaling) model for WRF and evaluate the different performance of three kinds of CPUs in different aspects, and at last, based the analysis of the attributes of the three CPUs and the nonlinear speedup of WRF under different numbers of resources, simulations result are given to address the energy consumption of WRF under different environments. We hope our research can help us to enhance the scheduling method of parallel tasks.