Jiayin Lu,Chris H. Rycroft
Abstract:We present TriMe++, a multi-threaded software library designed for generating two-dimensional meshes for intricate geometric shapes using the Delaunay triangulation. Multi-threaded parallel computing is implemented throughout the meshing procedure, making it suitable for fast generation of large-scale meshes. Three iterative meshing algorithms are implemented: the DistMesh algorithm, the centroidal Voronoi diagram meshing, and a hybrid of the two. We compare the performance of the three meshing methods in TriMe++, and show that the hybrid method retains the advantages of the other two. The software library achieves significant parallel speedup when generating a large mesh with $10^6$ points. TriMe++ can handle complicated geometries and generates adaptive meshes of high quality.
Computational Geometry,Distributed, Parallel, and Cluster Computing,Numerical Analysis,Applied Physics,Computational Physics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to efficiently generate high - quality triangular meshes on complex two - dimensional geometric shapes. Specifically, the paper introduces TriMe++, a multi - threaded software library for generating two - dimensional triangular meshes of complex geometric shapes, using Delaunay triangulation technology. The main contribution of the paper lies in the implementation of multi - threaded parallel computing throughout the mesh generation process, making it possible to rapidly generate large - scale meshes. In addition, the paper compares the performance of three iterative mesh generation algorithms (the DistMesh algorithm, the centroidal Voronoi diagram mesh generation algorithm, and a hybrid method of the two) in TriMe++, and shows that the hybrid method retains the advantages of the other two methods.
### Key Problems and Solutions
1. **Problem**: How to efficiently generate high - quality triangular meshes on complex two - dimensional geometric shapes?
- **Solution**: The paper proposes TriMe++, a multi - threaded software library based on Delaunay triangulation. By implementing multi - threaded parallel computing, TriMe++ can achieve significant parallel acceleration in large - scale mesh generation. The paper also implements three iterative mesh generation algorithms and verifies the effectiveness of these algorithms through experiments.
2. **Problem**: How to handle complex geometric shapes and generate high - quality meshes with adaptive resolution?
- **Solution**: TriMe++ can handle complex geometric shapes and generate high - quality meshes with adaptive resolution. As mentioned in the paper, the software library achieves efficient mesh generation through multi - threaded parallel computing, especially when generating large - scale meshes containing 10^6 points.
3. **Problem**: How to select an appropriate mesh generation algorithm to balance computational efficiency and mesh quality?
- **Solution**: The paper compares three mesh generation algorithms (the DistMesh algorithm, the centroidal Voronoi diagram mesh generation algorithm, and a hybrid method of the two) and analyzes their performance in terms of parallelization and mesh quality through experiments. The results show that the hybrid method has a low serial computation time while maintaining high - quality meshes, similar to the DistMesh algorithm.
### Main Contributions
- **Multi - threaded Parallel Computing**: TriMe++ implements multi - threaded parallel computing, significantly improving the efficiency of large - scale mesh generation.
- **Multiple Mesh Generation Algorithms**: The paper implements three iterative mesh generation algorithms and verifies their performance through experiments.
- **High - Quality Mesh Generation**: TriMe++ can handle complex geometric shapes and generate high - quality meshes with adaptive resolution.
- **Hybrid Method**: The paper proposes a hybrid method that combines the DistMesh and CVD mesh generation methods, which not only maintains high - quality meshes but also reduces the computation time.
### Summary
This paper solves the problem of efficiently generating high - quality triangular meshes on complex geometric shapes by introducing TriMe++. Through the implementation of multi - threaded parallel computing and multiple mesh generation algorithms, the paper not only improves the efficiency of mesh generation but also ensures the quality of the meshes. The proposed hybrid method further optimizes the mesh generation process, making it have a lower computation time while maintaining high quality.