CRTC top img
 

 

Parallel Constrained Delaunay Mesh Generation

 


 

 

Description

 

Delaunay refinement is a widely used method for the construction of guaranteed quality triangular and tetrahedral meshes. We present an algorithm and a software for the parallel constrained Delaunay mesh generation in two dimensions. Our approach is based on the decomposition of the original mesh generation problem into N smaller subproblems which are meshed in parallel. The parallel algorithm is asynchronous with small messages which can be aggregated and exhibits low communication costs. On a heterogeneous cluster of more than 100 processors our implementation can generate over one billion triangles in less than 3 minutes, while the single-node performance is comparable to that of the fastest to our knowledge sequential guaranteed quality Delaunay meshing library (Triangle). See Figure 1 and Figure 2 for some results.

 


 

 

Results

 

Figure 1 shows the Chesapeake bay model decomposed into 1024 subdomains, which are mapped onto 8 processes. The assignment of subdomains to processes is shown with different colors.

 

click to see picture in original size

Figure 1. The Chesapeake Bay Model

 

 

Figure 2 demonstrates the scaled speedup measurements. The number of triangles is approximately 10P, i.e., for 2 processors 20M, and for 144 processors about 1.4B.

 

click to see picture in original size

Figure 2. Scaled speedup measurements

  •  


 

Software

 

The following software is used in this project.   

  •     
          - meDDec
  •     
          - Metis
  •  


 

 

 

 

Related Papers

 

1. Parallel 2D Constrained Delaunay Mesh Generation

Andrey Chernikov and Nikos Chrisochoides.
ACM Transactions of Mathematical Software (accepted), February, 2007.

  [PDF]          [BibTex]     


 

2.  Multigrain Parallel Delaunay Mesh Generation: Challenges and Opportunities for Multithreaded Architectures

Christos D. Antonopoulos, Xiaoning Ding, Andrey Chernikov, Filip Blagojevic, Dimitrios S. Nikolopoulos and Nikos Chrisochoides.
In Proceedings of the 19th ACM International Conference on Supercomputing (ICS'2005), pp367-376. Cambridge, Massachusetts. June 2005.

  [PDF]          [BibTex]     


 

3.  Load Balancing Framework for Adaptive and Asynchronous Applications

Kevin Barker, Andrey Chernikov, Nikos Chrisochoides and Keshav Pingali.
IEEE Transactions on Parallel and Distributed Systems, Vol 14, No 12, pp 183-192, Feb. 2004.

  [PDF]          [BibTex]     


 

4.  A Multigrain Delaunay Mesh Generation Method for Multicore SMT-based Architectures

C. Antonopoulos, F. Blagojevic, A. Chernikov, D. Nikolopoulos and N. Chrisochoides.
Journal of Parallel and Distributed Systems, August (under revision) 2006.

  [PDF]          [BibTex]     


 

5.  Algorithm Software and Hardware Optimizations for Delaunay Mesh Generation on Simultaneous Multithreaded Architectures

C. Antonopoulos, F. Blagojevic, A. Chernikov, D. Nikolopoulos and N. Chrisochoides.
Journal of Parallel and Distributed Systems, August (under revision) 2006.

  [PDF]          [BibTex]     


 

6. Experience with Memory Allocators for Parallel Mesh Generation on Multicore Architectures

Andrey Chernikov, Christos Antonopoulos, Scott Schneider, Dimitrios Nikolopoulos and Nikos Chrisochoides.
To appear in the proceedngs of the 10th ISGG Conference on Numerical Grid Generation, October, 2007.

  [PDF]          [BibTex]     


 

7.  Parallel Out-of-Core Delaunay Refinement

Andriy Kot, Andrey Chernikov and Nikos Chrisochoides.
In Proceedings of Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, Sofia, Bulgaria, September 2005.

  [PDF]          [BibTex]     


 

For questions / comments, please contact Andrey Chernikov.