A GPU Solver for Sparse Generalized Eigenvalue Problems with Symmetric Complex-Valued Matrices Obtained Using Higher-Order FEM
Abstract
The paper discusses a fast implementation of the stabilized locally optimal block preconditioned conjugate gradient (sLOBPCG) method, using a hierarchical multilevel preconditioner to solve nonHermitian sparse generalized eigenvalue problems with large symmetric complex-valued matrices obtained using the higher-order finite-element method (FEM), applied to the analysis of a microwave resonator. The resonant frequencies of the low-order modes are the eigenvalues of the smallest real part of a complex symmetric (though non-Hermitian) matrix pencil. These type of pencils arise in the FEM analysis of resonant cavities loaded with a lossy material. To accelerate the computations, graphics processing units (GPU, NVIDIA Pascal P100) were used. Single and dual-GPU variants are considered and a GPU-memorysaving implementation is proposed. An efficient sliced ELLR-T sparse matrix storage format was used and operations were performed on blocks of vectors for best performance on a GPU. As a result, significant speedups (exceeding a factor of six in some computational scenarios) were achieved over the reference parallel implementation using a multicore central processing unit (CPU, Intel Xeon E5-2680 v3, twelve cores). These results indicate that the solution of generalized eigenproblems needs much more GPU memory than iterative techniques when solving a sparse system of equations, and also requires a second GPU to store some data structures in order to reduce the footprint, even for a moderately large systems
Citations
-
5
CrossRef
-
0
Web of Science
-
5
Scopus
Authors (2)
Cite as
Full text
- Publication version
- Accepted or Published Version
- License
- Copyright (2018 IEEE)
Keywords
Details
- Category:
- Articles
- Type:
- artykuł w czasopiśmie wyróżnionym w JCR
- Published in:
-
IEEE Access
no. 6,
pages 69826 - 69834,
ISSN: 2169-3536 - Language:
- English
- Publication year:
- 2018
- Bibliographic description:
- Dziekoński A., Mrozowski M.: A GPU Solver for Sparse Generalized Eigenvalue Problems with Symmetric Complex-Valued Matrices Obtained Using Higher-Order FEM// IEEE Access. -Vol. 6, (2018), s.69826-69834
- DOI:
- Digital Object Identifier (open in new tab) 10.1109/access.2018.2871219
- Bibliography: test
-
- M. Wang, H. Klie, M. Parashar, and H. Sudan, ''Solving sparse linear systems on NVIDIA Tesla GPUs,'' in Proc. Int. Conf. Comput. Sci., Berlin, Germany: Springer, 2009, pp. 864-873. open in new tab
- S. Georgescu and H. Okuda, ''Conjugate gradients on multiple GPUs,'' Int. J. Numer. Methods Fluids, vol. 64, nos. 10-12, pp. 1254-1273, 2010. open in new tab
- A. Dziekonski, A. Lamecki, and M. Mrozowski, ''GPU acceleration of multilevel solvers for analysis of microwave components with finite ele- ment method,'' IEEE Microw. Wireless Compon. Lett., vol. 21, no. 1, pp. 1-3, Jan. 2011. open in new tab
- A. Dziekonski, A. Lamecki, and M. Mrozowski, ''Tuning a hybrid GPU- CPU V-cycle multilevel preconditioner for solving large real and com- plex systems of FEM equations,'' IEEE Antennas Wireless Propag. Lett., vol. 10, pp. 619-622, 2011. open in new tab
- R. Helfenstein and J. Koko, ''Parallel preconditioned conjugate gradi- ent algorithm on GPU,'' J. Comput. Appl. Math., vol. 236, no. 15, pp. 3584-3590, 2012. open in new tab
- R. Li and Y. Saad, ''GPU-accelerated preconditioned iterative linear solvers,'' J. Supercomput., vol. 63, no. 2, pp. 443-466, 2013. open in new tab
- L. Z. Khodja, R. Couturier, A. Giersch, and J. M. Bahi, ''Parallel sparse linear solver with GMRES method using minimization techniques of communications for GPU clusters,'' J. Supercomput., vol. 69, no. 1, pp. 200-224, Jul. 2014.
- A. Dziekonski, P. Sypek, A. Lamecki, and M. Mrozowski, ''Communi- cation and load balancing optimization for finite element electromagnetic simulations using multi-GPU workstation,'' IEEE Trans. Microw. Theory Techn., vol. 65, no. 8, pp. 2661-2671, Aug. 2017. open in new tab
- A. Dziekonski and M. Mrozowski, ''Block conjugate-gradient method with multilevel preconditioning and GPU acceleration for FEM problems in electromagnetics,'' IEEE Antennas Wireless Propag. Lett., vol. 17, no. 6, pp. 1039-1042, Jun. 2018. open in new tab
- J. L. Aurentz, V. Kalantzis, and Y. Saad, ''Cucheb: A GPU implementation of the filtered lanczos procedure,'' Comput. Phys. Commun., vol. 220, pp. 332-340, Nov. 2017. open in new tab
- H. Anzt, S. Tomov, and J. Dongarra, ''On the performance and energy efficiency of sparse linear algebra on GPUs,'' Int. J. High Perform. Comput. Appl., vol. 31, no. 5, pp. 375-390, 2017. open in new tab
- H. Anzt, S. Tomov, and J. Dongarra, ''Accelerating the LOBPCG method on GPUs using a blocked sparse matrix vector product,'' in Proc. Symp. High Perform. Comput., 2015, pp. 75-82. open in new tab
- M. Kreutzer et al., ''Performance engineering and energy efficiency of building blocks for large, sparse eigenvalue computations on heteroge- neous supercomputers,'' in Software for Exascale Computing-SPPEXA. Cham, Switzerland: Springer, 2016, pp. 317-338. open in new tab
- M. Kreutzer et al., ''Chebyshev filter diagonalization on modern manycore processors and GPGPUs,'' in Proc. Int. Conf. High Perform. Comput.. open in new tab
- Cham, Switzerland: Springer, 2018, pp. 329-349.
- W. Rodrigues, A. Pecchia, M. A. der Maur, and A. Di Carlo, ''A compre- hensive study of popular eigenvalue methods employed for quantum cal- culation of energy eigenstates in nanostructures using GPUs,'' J. Comput. Electron., vol. 14, no. 2, pp. 593-603, 2015. open in new tab
- A. V. Knyazev, ''Toward the optimal preconditioned eigensolver: Locally optimal block preconditioned conjugate gradient method,'' SIAM J. Sci. Comput., vol. 23, no. 2, pp. 517-541, 2001. open in new tab
- M. Rewienski, A. Dziekonski, A. Lamecki, and M. Mrozowski, ''A stabi- lized complex LOBPCG eigensolver for the analysis of moderately lossy EM structures,'' IEEE Microw. Wireless Compon. Lett., vol. 28, no. 1, pp. 7-9, Jan. 2018. open in new tab
- A. Dziekonski, M. Rewienski, P. Sypek, A. Lamecki, and M. Mrozowski, ''GPU-accelerated LOBPCG method with inexact null-space filtering for solving generalized eigenvalue problems in computational electromagnet- ics analysis with higher-order FEM,'' Commun. Comput. Phys., vol. 22, no. 4, pp. 997-1014, 2017. open in new tab
- P. Ingelstrom, ''A new set of H (curl)-conforming hierarchical basis func- tions for tetrahedral meshes,'' IEEE Trans. Microw. Theory Techn., vol. 54, no. 1, pp. 106-114, Jan. 2006. open in new tab
- J. Rubio, J. Arroyo, and J. Zapata, ''Analysis of passive microwave circuits by using a hybrid 2-D and 3-D finite-element mode-matching method,'' IEEE Trans. Microw. Theory Techn., vol. 47, no. 9, pp. 1746-1749, Sep. 1999. open in new tab
- Z. Bai, J. Demmel, J. Dongarra, A. Ruhe, and H. van der Vorst, Templates for the Solution of Algebraic Eigenvalue Problems: A Practical Guide. Philadelphia, PA, USA: SIAM, 2000. open in new tab
- P. Arbenz and R. Geus, ''Multilevel preconditioned iterative eigensolvers for Maxwell eigenvalue problems,'' Appl. Numer. Math., vol. 54, no. 2, pp. 107-121, 2005. open in new tab
- A. Lamecki, L. Balewski, and M. Mrozowski, ''An efficient framework for fast computer aided design of microwave circuits based on the higher-order 3D finite-element method,'' Radioengineering, vol. 23, no. 4, pp. 970-978, 2014. open in new tab
- S. Filippone, V. Cardellini, D. Barbieri, and A. Fanfarillo, ''Sparse matrix- vector multiplication on GPGPUs,'' ACM Trans. Math. Softw., vol. 43, no. 4, 2017, Art. no. 30. open in new tab
- H. Anzt, S. Tomov, and J. Dongarra, ''Implementing a sparse matrix vector product for the SELL-C/SELL-C-σ formats on NVIDIA GPUs,'' Innov. open in new tab
- M. Clark, A. Strelchenko, A. Vaquero, M. Wagner, and E. Weinberg. (2017). ''Pushing memory bandwidth limitations through efficient imple- mentations of block-Krylov space solvers on GPUs.'' [Online]. Available: https://arxiv.org/abs/1710.09745 open in new tab
- M. Kreutzer, G. Hager, G. Wellein, H. Fehske, and A. R. Bishop, ''A uni- fied sparse matrix data format for efficient general sparse matrix-vector multiplication on modern processors with wide SIMD units,'' SIAM J. Sci. Comput., vol. 36, no. 5, pp. C401-C423, 2014. open in new tab
- A. Dziekonski, A. Lamecki, and M. Mrozowski, ''A memory efficient and fast sparse matrix vector product on a GPU,'' Prog. Electromagn. Res., vol. 16, pp. 49-63, 2011. open in new tab
- F. Vázquez, G. Ortega, J. J. Fernández, and E. M. Garzón, ''Improving the performance of the sparse matrix vector product with GPUs,'' in Proc. 10th IEEE Int. Conf. Comput. Inf. Technol., Jun./Jul. 2010, pp. 1146-1151. open in new tab
- A. Dziekonski and M. Mrozowski, ''GPU acceleration of block Krylov methods for FEM problems in electromagnetics,'' in Proc. IEEE MTT-S Int. Conf. Numer. Electromagn. Multiphys. Modeling Optim. RF, Microw., Terahertz Appl. (NEMO), May 2017, pp. 278-280. open in new tab
- S. Papantonis and S. Lucyszyn, ''Lossy spherical cavity resonators for stress-testing arbitrary 3D eigenmode solvers,'' Prog. Electromagn. Res., vol. 151, pp. 151-167, 2015. open in new tab
- J. Schöberl, ''NETGEN an advancing front 2D/3D-mesh generator based on abstract rules,'' Comput. Vis. Sci., vol. 1, no. 1, pp. 41-52, 1997. open in new tab
- ADAM DZIEKONSKI received the M.S.E.E. and Ph.D. degrees (Hons.) in microwave engineering from the Gdańsk University of Technology, Gdańsk, Poland, in 2009 and 2015, respectively. His current research interests include computational electromagnetics (mainly focused on parallelizing computa- tions on graphics processing and central processing units). He was a recipient of the Domestic Grant for Young Scientists from the Foundation for Polish Science in 2012 and 2013. He was also a recipient of the Prime Minister's Award for his Ph.D. thesis in 2016. open in new tab
- MICHAL MROZOWSKI (F'08) received the M.Sc. and Ph.D. degrees (Hons.) from the Gdańsk University of Technology in 1983 and 1990, respectively. In 1986, he joined the Faculty of Electronics, Gdańsk University of Technology, where he is currently a Full Professor, the Head of the Department of Microwave and Antenna Engineering, and the Director of the Center of Excellence for Wireless Communication Engineering. open in new tab
- Verified by:
- Gdańsk University of Technology
seen 129 times