Abstract
Load balancing plays an important role in parallel numerical simulations. State-of-the-art libraries addressing this problem base on vertex exchange heuristics that are embedded in a multilevel scheme. However, these are hard to parallelize due to their sequential nature. Furthermore, libraries like Metis and Jostle focus on a small edge-cut and cannot obey constraints like connectivity and straight partition boundaries, which are important for some numerical solvers.
In this paper we present an alternative approach to balance the load in parallel adaptive finite element simulations. We compute a distribution that is based on solutions of linear equations. Integrated into a learning framework, we obtain a heuristic that contains a high degree of parallelism and computes well shaped connected partitions. Furthermore, our experiments indicate that we can find solutions that are comparable to those of the two state-of-the-art libraries Metis and Jostle also regarding the classic metrics like edge-cut and boundary length.
This work is supported by the German Science Foundation (DFG) project SFB-376 and by DFG Research Training Group GK-693.
Chapter PDF
Similar content being viewed by others
References
Fox, G., Williams, R., Messina, P.: Parallel Computing Works! Morgan Kaufmann, San Francisco (1994)
Karypis, G., Kumar, V.: MeTis: A Software Package for Partitioning Unstrctured Graphs, Partitioning Meshes, [...], Version 4.0 (1998)
Walshaw, C.: The parallel JOSTLE library user guide: Version 3.0 (2002)
Schamberger, S.: Graph partitioning with the Party library: Helpful-sets in practice. In: Comp. Arch. and High Perf. Comp., SBAC-PAD 2004, pp. 198–205 (2004)
Hendrickson, B., Leland, R.: A multi-level algorithm for partitioning graphs. In: Supercomputing 1995 (1995)
Kernighan, B.W., Lin, S.: An efficient heuristic for partitioning graphs. Bell Systems Technical Journal 49, 291–308 (1970)
Hromkovic, J., Monien, B.: The bisection problem for graphs of degree 4. In: Tarlecki, A. (ed.) MFCS 1991. LNCS, vol. 520, pp. 211–220. Springer, Heidelberg (1991)
Schloegel, K., Karypis, G., Kumar, V.: Multilevel diffusion schemes for repartitioning of adaptive meshes. J. Par. Dist. Comp. 47(2), 109–124 (1997)
Walshaw, C., Cross, M.: Parallel optimisation algorithms for multilevel mesh partitioning. J. Parallel Computing 26(12), 1635–1660 (2000)
Hendrickson, B.: Graph partitioning and parallel solvers: Has the emperor no clothes? In: Ferreira, A., Rolim, J.D.P., Teng, S.-H. (eds.) IRREGULAR 1998. LNCS, vol. 1457, pp. 218–225. Springer, Heidelberg (1998)
Oliker, L., Biswas, R.: PLUM: Parallel load balancing for adaptive unstructured meshes. J. Par. Dist. Comp. 52(2), 150–177 (1998)
Diekmann, R., Preis, R., Schlimbach, F., Walshaw, C.: Shape-opt. mesh part. and load bal. for par. adap. FEM. J. Parallel Computing 26, 1555–1581 (2000)
Schamberger, S.: On partitioning FEM graphs using diffusion. In: HPGC, Intern. Parallel and Distributed Processing Symposium, IPDPS 2004(CD), p. 277 (2004)
Elsässer, R., Monien, B., Preis, R.: Diffusion schemes for load balancing on heterogeneous networks. Theory of Computing Systems 35, 305–320 (2002)
Schamberger, S.: A shape optimizing load distribution heuristic for parallel adaptive FEM computations. Accepted at PACT 2005 (2005)
Hu, Y.F., Blake, R.F.: An improved diffusion algorithm for dynamic load balancing. Parallel Computing 25(4), 417–444 (1999)
Marquardt, O., Schamberger, S.: Open benchmarks for load balancing heuristics in parallel adaptive finite element computations. Accepted at PDPTA 2005 (2005)
Schamberger, S.: http://www.upb.de/cs/schaum/benchmark.html
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2005 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Meyerhenke, H., Schamberger, S. (2005). Balancing Parallel Adaptive FEM Computations by Solving Systems of Linear Equations. In: Cunha, J.C., Medeiros, P.D. (eds) Euro-Par 2005 Parallel Processing. Euro-Par 2005. Lecture Notes in Computer Science, vol 3648. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11549468_26
Download citation
DOI: https://doi.org/10.1007/11549468_26
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-28700-1
Online ISBN: 978-3-540-31925-2
eBook Packages: Computer ScienceComputer Science (R0)