Title :
CPOP: Component Design and Parallelization towards POP Ocean Model Based on ESMF
Author :
Yang, Hailong ; Zhao, Qi ; Luan, Zhongzhi ; Qian, Depei
Author_Institution :
Dept. of Comput. Sci. & Eng., Beihang Univ., Beijing, China
Abstract :
Nowadays, as the research of climate model developing to simulate the interactions between multiple physical procedures and demand high resolution results, the complexity of the climate model has increased dramatically. The traditional tight-coupling approach applied in model design has expressed tremendous burden on the code maintenance and update due to the complicated relationships between function calls and data sharing in global scope. Such an obstacle has severely hindered the rapid innovation and development of climate models, especially when involving interdisciplinary efforts and collaborations among multiple research institutes. In this study, we present our approach to redesign POP into component based CPOP and apply hybrid parallelization at component level, both of which aim at solving the mentioned limitations.
Keywords :
climatology; geophysics computing; object-oriented programming; oceanography; parallel processing; ESMF; POP ocean model; climate model complexity; code maintenance; component based CPOP; component design; component level; data sharing; function call; hybrid parallelization; parallel ocean program; Computational modeling; Mathematical model; Meteorology; Oceans; Parallel processing; Positron emission tomography; Software; Component Encapsulation; ESMF; Hybrid Parallelization; POP;
Conference_Titel :
Parallel Processing Workshops (ICPPW), 2012 41st International Conference on
Conference_Location :
Pittsburgh, PA
Print_ISBN :
978-1-4673-2509-7
DOI :
10.1109/ICPPW.2012.86