Multi-GPU Acceleration for Finite Element Analysis in Structural Mechanics
Source abstract
Abstract
This work evaluates the computing performance of finite element analysis in structural mechanics using modern multi-GPU systems. We can avoid the usual memory limitations when using one GPU device for many-core computing using multiple GPUs for scientific computing. We use a GPU-awareness MPI approach implementing a suitable smoothed aggregation multigrid for preconditioning an iterative distributed conjugate gradient solver for GPU computing. We evaluate the performance and scalability of different models, problem sizes, and computing resources. We take an efficient multi-core implementation as the reference to assess the computing performance of the numerical results. The numerical results show the advantages and limitations of using distributed many-core architectures to address structural mechanics problems.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Topology Optimization in Engineering · Robotic Mechanisms and Dynamics · Advanced Numerical Methods in Computational Mathematics
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Escuela Técnica Superior de Ingeniería Industrial, Universidad Politécnica de Cartagena, Campus Muralla del Mar, 30202 Cartagena, Spain
Facultad de Informática, Universidad de Murcia, 30100 Murcia, Spain
Funding
AEI/FEDER
References (60)
Show source-provided references
1. Structural optimization complexity: What has Moore’s law done for us?10.1007/s00158-004-0415-y ↗
2. Moore’s Law: Past, present and future10.1109/6.591665 ↗
3. A View of the Parallel Computing Landscape10.1145/1562764.1562783 ↗
4. Wall, D. (1991, January 8–11). Limits of instruction-level parallelism. Proceedings of the Fourth International Conference on Architectural Support for Programming Languages and Operating Systems, Santa Clara, CA, USA.10.1145/106972.106991 ↗
5. Hitting the memory wall: Implications of the obvious10.1145/216585.216588 ↗
6. Asanovic, K., Bodik, R., Catanzaro, B.C., Gebis, J.J., Husbands, P., Keutzer, K., Patterson, D.A., Plishker, W.L., Shalf, J., and Williams, S.W. (2006). The Landscape of Parallel Computing Research: A View from Berkeley, UC Berkeley. Technical report.
7. Kahle, J., Moreno, J., and Dreps, D. (2019, January 17–21). Summit and Sierra: Designing AI/HPC Supercomputers. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.
8. The GPU Computing Era10.1109/MM.2010.41 ↗
9. Graphics processing unit (GPU) programming strategies and trends in GPU computing10.1016/j.jpdc.2012.04.003 ↗
10. Noaje, G., Krajecki, M., and Jaillet, C. (2010, January 26–28). MultiGPU computing using MPI or OpenMP. Proceedings of the International Conference on Intelligent Computer Communication and Processing, Cluj-Napoca, Romania.10.1109/ICCP.2010.5606414 ↗
11. Cheik-Ahamed, A.K., and Magoulès, F. (2016, January 12–14). GPU accelerated substructuring methods for sparse linear systems. Proceedings of the IEEE Int. Conf. on High Performance Computing and Communications (HPCC), Paris, France.
12. Multi-GPU parallel acceleration scheme for meshfree peridynamic simulations10.1016/j.tafmec.2024.104401 ↗
13. MDScale: Scalable multi-GPU bonded and short-range molecular dynamics10.1016/j.jpdc.2021.07.006 ↗
14. DUGKS-GPU: An efficient parallel GPU code for 3D turbulent flow simulations using Discrete Unified Gas Kinetic Scheme10.1016/j.cpc.2024.109216 ↗
15. Optimization of Large-Scale Sparse Matrix-Vector Multiplication on Multi-GPU Systems
16. Multi-GPU work sharing in a task-based dataflow programming model10.1016/j.future.2024.03.017 ↗
17. GPU Acceleration for FEM-Based Structural Analysis10.1007/s11831-013-9082-8 ↗
18. Kirk, D.B., and Hwu, W.m.W. (2013). Programming Massively Parallel Processors: A Hands-on Approach, Morgan Kaufmann. [2nd ed.].
19. MPI-based implementation of a PCG solver using an EBE architecture and preconditioner for implicit, 3-D finite element analysis10.1016/S0045-7949(00)00153-X ↗
20. Object-oriented programming of distributed iterative equation solvers10.1016/j.compstruc.2007.05.003 ↗
21. Enhancing the Performance of Conjugate Gradient Solvers on Graphic Processing Units10.1109/TMAG.2010.2081662 ↗
22. Parallel preconditioned conjugate gradient algorithm on GPU10.1016/j.cam.2011.04.025 ↗
23. GPU-accelerated preconditioned iterative linear solvers10.1007/s11227-012-0825-3 ↗
24. Exploring weak scalability for FEM calculations on a GPU-enhanced cluster10.1016/j.parco.2007.09.002 ↗
25. Bell, N., and Garland, M. (2008). Efficient Sparse Matrix-Vector Multiplication on CUDA, NVIDIA. NVIDIA Technical Report NVR-2008-004.
26. Cheik-Ahamed, A.K., and Magoulès, F. (2014, January 20–22). Parallel Sub-Structuring Methods for solving Sparse Linear Systems on a cluster of GPU. Proceedings of the IEEE International Conference on High Performance Computing and Communications (HPCC), Paris, France.
27. Efficient implementation of Jacobi iterative method for large sparse linear systems on graphic processing units10.1007/s11227-016-1701-3 ↗
28. Fine-grained GPU implementation of assembly-free iterative solver for finite element problems10.1016/j.compstruc.2015.05.010 ↗
29. Assembly of finite element methods on graphics processors10.1002/nme.2989 ↗
30. Finite element assembly strategies on multi-core and many-core architectures10.1002/fld.3648 ↗
31. Parallel Realization of the Element-by-Element FEM Technique by CUDA10.1109/TMAG.2011.2175905 ↗
32. A Parallel Node-based Solution Scheme for Implicit Finite Element Method Using GPU10.1016/j.proeng.2013.08.022 ↗
33. Efficient Matrix-Free GPU implementation of Fixed Grid Finite Element Analysis10.1016/j.finel.2015.06.005 ↗
34. Efficient generation of large-scale pareto-optimal topologies10.1007/s00158-012-0807-3 ↗
35. Toward GPU accelerated topology optimization on unstructured meshes10.1007/s00158-013-0920-y ↗
36. Ewing, R. (1987). Algebraic Multigrid. Multigrid Methods, Society for Industrial and Applied Mathematics (SIAM), 3600 University City Science Center. Chapter 4.
37. Multi-level iterative technique and aggregation concept with semi-analytical preconditioning for solving boundary-value problem10.1002/cnm.1640090804 ↗
38. A GPU accelerated aggregation algebraic multigrid method10.1016/j.camwa.2014.08.022 ↗
39. Analysis and Comparison of Relaxation Schemes in Robust Multigrid and Preconditioned Conjugate Gradient Methods10.1007/BFb0069941 ↗
40. Tatebe, O. (1993, January 4–9). The multigrid preconditioned conjugate gradient method. Proceedings of the Sixth Copper Mountain Conference on Multigrid Methods, Copper Mountain, CO, USA.
41. A review of algebraic multigrid10.1016/S0377-0427(00)00516-1 ↗
42. A Real-Time Multigrid Finite Hexahedra Method for Elasticity Simulation using CUDA10.1016/j.simpat.2010.11.005 ↗
43. GPU acceleration for evolutionary topology optimization of continuum structures using isosurfaces10.1016/j.compstruc.2016.10.018 ↗
44. Efficient topology optimization using GPU computing with multilevel granularity10.1016/j.advengsoft.2017.01.009 ↗
45. A System for High-Resolution Topology Optimization10.1109/TVCG.2015.2502588 ↗
46. Architecting the finite element method pipeline for the GPU10.1016/j.cam.2013.09.001 ↗
47. Accelerating algebraic multigrid solvers on NVIDIA GPUs10.1016/j.camwa.2015.07.005 ↗
48. Karypis, G., and Schloegel, K. (2013). ParMeTis: Parallel Graph Partitioning and Sparse Matrix Ordering Library, Version 4.0, University of Minnesota. Technical report.
49. AMGCL: An Efficient, Flexible, and Extensible Algebraic Multigrid Implementation10.1134/S1995080219050056 ↗
50. Programming CUDA and OpenCL: A Case Study Using Modern C++ Libraries10.1137/120903683 ↗
51. Kindratenko, V. (2014). Solving Ordinary Differential Equations on GPUs. Numerical Computations with GPUs, Springer International Publishing.10.1007/978-3-319-06548-9 ↗
52. Szuppe, J. (2016, January 19–21). Boost.Compute: A parallel computing library for C++ based on OpenCL. Proceedings of the 4th International Workshop on OpenCL (IWOCL 16), Vienna, Austria.10.1145/2909437.2909454 ↗
53. Acceleration of Convergence of a Two Level Algorithm by Smooth Transfer Operators10.21136/AM.1992.104509 ↗
54. Algebraic multigrid by smoothed aggregation for second and fourth order elliptic problems10.1007/BF02238511 ↗
55. Tuminaro, R., and Tong, C. (2010, January 13–19). Parallel Smoothed Aggregation Multigrid: Aggregation Strategies on Massively Parallel Machines. Proceedings of the International ACM/IEEE Conference on Supercomputing, Dallas, TX, USA.
56. Exposing Fine-Grained Parallelism in Algebraic Multigrid Methods10.1137/110838844 ↗
57. Parallel Preconditioning with Sparse Approximate Inverses10.1137/S1064827594276552 ↗
58. Parallel solution techniques in computational structural mechanics10.1016/S0045-7825(97)00028-5 ↗
59. Non-overlapping Schwarz methods with optimized transmission conditions for the Helmholtz equation10.1016/j.cma.2004.05.004 ↗
60. Multilevel k-way partitioning scheme for irregular graphs10.1006/jpdc.1997.1404 ↗