This discussion paper is a preprint. It is a manuscript under review for the journal Geoscientific Model Development (GMD).
PatCC1: an Efficient Parallel Triangulation Algorithm for Spherical and Planar Grids with Commonality and Parallel Consistency
Haoyu Yang1,Li Liu1,2,Cheng Zhang1,2,Ruizhe Li1,2,Chao Sun1,Xinzhu Yu1,Hao Yu1,Zhiyuan Zhang3,and Bin Wang1,2,4Haoyu Yang et al. Haoyu Yang1,Li Liu1,2,Cheng Zhang1,2,Ruizhe Li1,2,Chao Sun1,Xinzhu Yu1,Hao Yu1,Zhiyuan Zhang3,and Bin Wang1,2,4
1Ministry of Education Key Laboratory for Earth System Modeling, Department of Earth System Science, Tsinghua University, Beijing, China
2Joint Center for Global Change Studies (JCGCS), Beijing, China
3Hydro-Meteorological Center of Navy China, Beijing, China
4State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
1Ministry of Education Key Laboratory for Earth System Modeling, Department of Earth System Science, Tsinghua University, Beijing, China
2Joint Center for Global Change Studies (JCGCS), Beijing, China
3Hydro-Meteorological Center of Navy China, Beijing, China
4State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Received: 11 Nov 2018 – Accepted for review: 29 Jan 2019 – Discussion started: 30 Jan 2019
Abstract. Graphs are commonly gridded by triangulation; i.e., the generation of a set of triangles for the points of the graph. This technique can also be used in a coupler to improve the commonality of data interpolation between different horizontal model grids. This paper proposes a new parallel triangulation algorithm, PatCC1 (Parallel triangulation algorithm with Commonality and parallel Consistency, version 1), for spherical and planar grids. Experimental evaluation results demonstrate the efficient parallelization of PatCC1 using a hybrid of MPI and OpenMP. They also show PatCC1 to have greater commonality than existing parallel triangulation algorithms (i.e., it is capable of handling more types of model grids) and that it guarantees parallel consistency (i.e., it achieves exactly the same triangulation result under different parallel settings).