Open AccessMesh Coarsening for Fast Simulation on Low Resource Machine
Author(s) -
Guoan Luo,
Zhiguo Zhang
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/790/1/012128
Subject(s) - computer science , t vertices , quadrilateral , laplacian smoothing , debugging , boundary (topology) , algorithm , parallel computing , mesh generation , polygon mesh , domain (mathematical analysis) , resource (disambiguation) , wire mesh , computational science , finite element method , mathematics , computer graphics (images) , operating system , materials science , engineering , structural engineering , composite material , mathematical analysis , computer network
Mesh-based application uses a mesh to discretely represent a real physical system and then calls appropriate algorithm over the mesh to solve equation for simulation. Mesh-based application is usually developed by domain programmers and it runs with very large mesh on super computer. It takes a long time to run with such a large mesh on low resource machine. In this paper, we propose an algorithm designed for coarsening two-dimensional quadrilateral mesh aiming to reduce the size of the mesh so that the application could run with the smaller coarse mesh on low resource machine in a short time. This makes it more convenient for domain programmers to develop and debug the application on low resource machine. The algorithm first generates cells of the coarse mesh by agglomerating cells of the original mesh. Then it constructs internal edges, boundary edges, nodes and all mappings of the coarse mesh. At last, the algorithm constructs datasets of the coarse mesh according to the datasets of the original mesh. The experimental result shows that it takes much less time to run application over the coarse mesh produced by our algorithm, and the error of fast simulation results is within the acceptable range.