ABSTRACT
In order to adapt various complex parallel environments, increase the utilization ratio of current software and hardware, as well as improve the overall parallel efficiency of parallel FEM, the Parallel Node-based Local Finite Element (PNLFE) System is presented. The PNLFE system separates the whole computing task to N unit tasks according to the number of mesh nodes N. For each mesh node, the unit task is to generate the corresponding local mesh, which meets conforming requirement, then compute the global stiffness matrix row of the node. After this is done, processors can immediately pre-process the stiffness matrix rows and obtain the numerical solution with the least communication compared with other known systems. Unit tasks are separated and small, thus the system is applicable to various parallel architectures.
- B. Glut and T. Jurczyk. Domain decomposition techniques for parallel generation of tetrahedral meshes. In ICCS '08: Proceedings of the 8th international conference on Computational Science, Part I, pages 641--650, Berlin, Heidelberg, 2008. Springer-Verlag. Google ScholarDigital Library
- W. Wang, G. Kosakowski, and O. Kolditz. A parallel finite element scheme for thermo-hydro-mechanical (THM) coupled problems in porous media. Comput. Geosci., 35(8):1631--1641, 2009. Google ScholarDigital Library
- N. Chrisochoides. Parallel mesh generation. Numerical Solution of Partial Differential Equations on Parallel Computers, 51, 2005. Eds. Are Magnus Bruaset, Petter Bjorstad, Aslak Tveito.Google Scholar
- L. Grigori, B. Philippe, A. Sameh, D. Tromeur-Dervout, and M. Vajtersic. Parallel matrix algorithms and applications. Parallel Computing, 34(6-8):293--295, 2008. Google ScholarDigital Library
- G. Yagawa. Node-by-node parallel finite elements: a virtually meshless method. International Journal for Numerical Methods in Engineering, 60(1):69--102, 2004.Google ScholarCross Ref
- Y. Nie, Y. Liu, Y. Gu, and X. Fan. Fast Searching Algorithm for Candidate Satellite-node Set in NLMG. CMES-Computer Modeling in Engineering & Sciences, 45(1):31--55,2009.Google Scholar
Index Terms
- A portable parallel finite element simulation system
Recommendations
Metrics for evaluation of parallel efficiency toward highly parallel processing
I studied novel performance metrics of parallel processing efficiency, determined simply by measuring timing data in a parallel process. The metrics, which consider load-imbalance, are called the parallel efficiency, load-balancing, and parallel ...
Performance study of the domain decomposition method with direct equation solver for parallel finite element analysis
AbstractA parallel performance of the domain decomposition method with directLDU algorithm of condensation and solution is studied. Typical subdomains arising after division of a square domain are considered, and operation count equations for all steps of ...
Massively parallel finite element programming
EuroMPI'10: Proceedings of the 17th European MPI users' group meeting conference on Recent advances in the message passing interfaceToday's large finite element simulations require parallel algorithms to scale on clusters with thousands or tens of thousands of processor cores. We present data structures and algorithms to take advantage of the power of high performance computers in ...
Comments