Article
| Title: | Resilient asynchronous primal Schur method (English) |
| Author: | Gbikpi-Benissan, Guillaume |
| Author: | Magoulès, Frédéric |
| Language: | English |
| Journal: | Applications of Mathematics |
| ISSN: | 0862-7940 (print) |
| ISSN: | 1572-9109 (online) |
| Volume: | 67 |
| Issue: | 6 |
| Year: | 2022 |
| Pages: | 679-704 |
| Summary lang: | English |
| . | |
| Category: | math |
| . | |
| Summary: | This paper introduces the application of asynchronous iterations theory within the framework of the primal Schur domain decomposition method. A suitable relaxation scheme is designed, whose asynchronous convergence is established under classical spectral radius conditions. For the usual case where local Schur complement matrices are not constructed, suitable splittings based only on explicitly generated matrices are provided. Numerical experiments are conducted on a supercomputer for both Poisson's and linear elasticity problems. The asynchronous Schur solver outperformed the classical conjugate-gradient-based one in case of computing node failures. (English) |
| Keyword: | asynchronous iterations |
| Keyword: | Schur complement method |
| Keyword: | domain decomposition method |
| Keyword: | parallel computing |
| MSC: | 49M20 |
| MSC: | 65F10 |
| MSC: | 65N55 |
| MSC: | 65Y05 |
| idZBL: | Zbl 07613019 |
| idMR: | MR4505700 |
| DOI: | 10.21136/AM.2022.0146-21 |
| . | |
| Date available: | 2022-10-31T13:24:57Z |
| Last updated: | 2023-11-24 |
| Stable URL: | http://hdl.handle.net/10338.dmlcz/151052 |
| . | |
| Reference: | [1] Amdahl, G. M.: Validity of the single processor approach to achieving large scale computing capabilities.Proceedings of the Conference AFIPS Joint Computer Conference, Atlantic City, New Jersey, USA ACM, New York (1967), 483-485. 10.1145/1465482.1465560 |
| Reference: | [2] Axelsson, O., Kolotilina, L.: Diagonally compensated reduction and related preconditioning methods.Numer. Linear Algebra Appl. 1 (1994), 155-177. Zbl 0837.65023, MR 1277800, 10.1002/nla.1680010207 |
| Reference: | [3] Bahi, J., Griepentrog, E., Miellou, J. C.: Parallel treatment of a class of differential-algebraic systems.SIAM J. Numer. Anal. 33 (1996), 1969-1980. Zbl 0859.65073, MR 1411858, 10.1137/S0036142993258105 |
| Reference: | [4] Baudet, G. M.: Asynchronous iterative methods for multiprocessors.J. Assoc. Comput. Mach. 25 (1978), 226-244. Zbl 0372.68015, MR 0494894, 10.1145/322063.322067 |
| Reference: | [5] Bertsekas, D. P., Tsitsiklis, J. N.: Parallel and Distributed Computation: Numerical Methods.Prentice-Hall, Englewood Cliffs (1989). Zbl 0743.65107, MR 3587745 |
| Reference: | [6] Bertsekas, D. P., Tsitsiklis, J. N.: Some aspects of parallel and distributed iterative algorithms: A survey.Automatica 27 (1991), 3-21. Zbl 0728.65041, MR 1087139, 10.1016/0005-1098(91)90003-K |
| Reference: | [7] Castel, M. J., Migallón, V., Penadés, J.: Convergence of non-stationary parallel multisplitting methods for Hermitian positive definite matrices.Math. Comput. 67 (1998), 209-220. Zbl 0895.65006, MR 1433264, 10.1090/S0025-5718-98-00893-X |
| Reference: | [8] Chazan, D., Miranker, W.: Chaotic relaxation.Linear Algebra Appl. 2 (1969), 199-222. Zbl 0225.65043, MR 0251888, 10.1016/0024-3795(69)90028-7 |
| Reference: | [9] Fan, K.: Topological proofs for certain theorems on matrices with non-negative elements.Monatsh. Math. 62 (1958), 219-237. Zbl 0081.25104, MR 0095856, 10.1007/BF01303967 |
| Reference: | [10] Fan, K.: Note on $M$-matrices.Q. J. Math., Oxf. II. Ser. 11 (1960), 43-49. Zbl 0104.01203, MR 0117242, 10.1093/qmath/11.1.43 |
| Reference: | [11] Frommer, A., Schwandt, H., Szyld, D. B.: Asynchronous weighted additive Schwarz methods.ETNA, Electron. Trans. Numer. Anal. 5 (1997), 48-61. Zbl 0890.65027, MR 1460886 |
| Reference: | [12] Frommer, A., Szyld, D. B.: $H$-splittings and two-stage iterative methods.Numer. Math. 63 (1992), 345-356. Zbl 0764.65018, MR 1186346, 10.1007/BF01385865 |
| Reference: | [13] Gbikpi-Benissan, G., Magoulès, F.: Protocol-free asynchronous iterations termination.Adv. Eng. Softw. 146 (2020), Article ID 102827, 9 pages. 10.1016/j.advengsoft.2020.102827 |
| Reference: | [14] Krivoshapko, S. N., Rynkovskaya, M.: Five types of ruled helical surfaces for helical conveyers, support anchors and screws.MATEC Web Conf. 95 (2017), Article ID 06002, 5 pages. 10.1051/matecconf/20179506002 |
| Reference: | [15] Magoulès, F., Gbikpi-Benissan, G.: Asynchronous parareal time discretization for partial differential equations.SIAM J. Sci. Comput. 40 (2018), C704--C725. Zbl 1404.65155, MR 3878314, 10.1137/17M1149225 |
| Reference: | [16] Magoulès, F., Gbikpi-Benissan, G.: Distributed convergence detection based on global residual error under asynchronous iterations.IEEE Trans. Parallel Distrib. Syst. 29 (2018), 819-829. 10.1109/TPDS.2017.2780856 |
| Reference: | [17] Magoulès, F., Gbikpi-Benissan, G.: JACK2: An MPI-based communication library with non-blocking synchronization for asynchronous iterations.Adv. Eng. Softw. 119 (2018), 116-133. 10.1016/j.advengsoft.2018.01.009 |
| Reference: | [18] Magoulès, F., Roux, F.-X., Salmon, S.: Optimal discrete transmission conditions for a nonoverlapping domain decomposition method for the Helmholtz equation.SIAM J. Sci. Comput. 25 (2004), 1497-1515. Zbl 1086.65118, MR 2087323, 10.1137/S1064827502415351 |
| Reference: | [19] Magoulès, F., Szyld, D. B., Venet, C.: Asynchronous optimized Schwarz methods with and without overlap.Numer. Math. 137 (2017), 199-227. Zbl 1382.65449, MR 3679933, 10.1007/s00211-017-0872-z |
| Reference: | [20] Magoulès, F., Venet, C.: Asynchronous iterative sub-structuring methods.Math. Comput. Simulation 145 (2018), 34-49. Zbl 07316164, MR 3725798, 10.1016/j.matcom.2016.05.009 |
| Reference: | [21] Schechter, S.: Relaxation methods for linear equations.Commun. Pure Appl. Math. 12 (1959), 313-335. Zbl 0096.09801, MR 0107361, 10.1002/cpa.3160120208 |
| Reference: | [22] Spiteri, P., Miellou, J.-C., Baz, D. El: Asynchronous Schwarz alternating method with flexible communication for the obstacle problem.Rés. Syst. Répartis Calculateurs Parallèles 13 (2001), 47-66. |
| Reference: | [23] Spiteri, P., Miellou, J.-C., Baz, D. El: Parallel asynchronous Schwarz and multisplitting methods for a nonlinear diffusion problem.Numer. Algorithms 33 (2003), 461-474. Zbl 1033.65085, MR 2005584, 10.1023/A:1025561332238 |
| Reference: | [24] Varga, R. S.: Matrix Iterative Analysis.Springer Series in Computational Mathematics 27. Springer, Berlin (2000). Zbl 0998.65505, MR 1753713, 10.1007/978-3-642-05156-2 |
| Reference: | [25] Yun, J. H., Kim, S. W.: Convergence of two-stage iterative methods using incomplete factorization.J. Comput. Appl. Math. 166 (2004), 565-580. Zbl 1089.65027, MR 2041199, 10.1016/j.cam.2003.09.041 |
| . |
Fulltext not available (moving wall 24 months)
Search
Partner of
