Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
143 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Scaling Molecular Dynamics with ab initio Accuracy to 149 Nanoseconds per Day (2410.22867v1)

Published 30 Oct 2024 in cs.DC

Abstract: Physical phenomena such as chemical reactions, bond breaking, and phase transition require molecular dynamics (MD) simulation with ab initio accuracy ranging from milliseconds to microseconds. However, previous state-of-the-art neural network based MD packages such as DeePMD-kit can only reach 4.7 nanoseconds per day on the Fugaku supercomputer. In this paper, we present a novel node-based parallelization scheme to reduce communication by 81%, then optimize the computationally intensive kernels with sve-gemm and mixed precision. Finally, we implement intra-node load balance to further improve the scalability. Numerical results on the Fugaku supercomputer show that our work has significantly improved the time-to-solution of the DeePMD-kit by a factor of 31.7x, reaching 149 nanoseconds per day on 12,000 computing nodes. This work has opened the door for millisecond simulation with ab initio accuracy within one week for the first time.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (46)
  1. R. Ainsworth, “The assessment of defects in structures of strain hardening material,” Engineering Fracture Mechanics, vol. 19, no. 4, pp. 633–642, 1984.
  2. J.-Y. Raty, F. Gygi, and G. Galli, “Growth of carbon nanotubes on metal nanoparticles: a microscopic mechanism from ab initio molecular dynamics simulations,” Physical review letters, vol. 95, no. 9, p. 096103, 2005.
  3. R. Callender and R. B. Dyer, “Advances in time-resolved approaches to characterize the dynamical nature of enzymatic catalysis,” Chemical reviews, vol. 106, no. 8, pp. 3031–3042, 2006.
  4. D. E. Shaw, R. O. Dror, J. K. Salmon, J. Grossman, K. M. Mackenzie, J. A. Bank, C. Young, M. M. Deneroff, B. Batson, K. J. Bowers et al., “Millisecond-scale molecular dynamics simulations on anton,” in Proceedings of the conference on high performance computing networking, storage and analysis, 2009, pp. 1–11.
  5. R. Car and M. Parrinello, “Unified approach for molecular dynamics and density-functional theory,” Physical review letters, vol. 55, no. 22, p. 2471, 1985.
  6. W. Jia, Z. Cao, L. Wang, J. Fu, X. Chi, W. Gao, and L.-W. Wang, “The analysis of a plane wave pseudopotential density functional theory code on a gpu machine,” Computer Physics Communications, vol. 184, no. 1, pp. 9–18, 2013.
  7. W. Jia, J. Fu, Z. Cao, L. Wang, X. Chi, W. Gao, and L.-W. Wang, “Fast plane wave density functional theory molecular dynamics calculations on multi-gpu machines,” Journal of Computational Physics, vol. 251, pp. 102–115, 2013. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S002199911300329X
  8. Y.-J. Yan, H.-B. Li, T. Zhao, L.-W. Wang, L. Shi, T. Liu, G.-M. Tan, W.-L. Jia, and N.-H. Sun, “10-million atoms simulation of first-principle package ls3df,” Journal of Computer Science and Technology, vol. 39, no. 1, pp. 45–62, 2024. [Online]. Available: https://jcst.ict.ac.cn/en/article/doi/10.1007/s11390-023-3011-6
  9. L. Wang, Y. Wu, W. Jia, W. Gao, X. Chi, and L.-W. Wang, “Large scale plane wave pseudopotential density functional theory calculations on gpu clusters,” in SC ’11: Proceedings of 2011 International Conference for High Performance Computing, Networking, Storage and Analysis, 2011, pp. 1–10.
  10. J. Behler and M. Parrinello, “Generalized neural-network representation of high-dimensional potential-energy surfaces,” Physical review letters, vol. 98, no. 14, p. 146401, 2007.
  11. A. Thompson, L. Swiler, C. Trott, S. Foiles, and G. Tucker, “Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentials,” Journal of Computational Physics, vol. 285, pp. 316–330, 2015. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0021999114008353
  12. K. Lee, D. Yoo, W. Jeong, and S. Han, “Simple-nn: An efficient package for training and executing neural-network interatomic potentials,” Computer Physics Communications, vol. 242, pp. 95–103, 2019.
  13. J. Behler, “Representing potential energy surfaces by high-dimensional neural network potentials,” Journal of Physics: Condensed Matter, vol. 26, no. 18, p. 183001, apr 2014. [Online]. Available: https://doi.org/10.1088/0953-8984/26/18/183001
  14. J. Behler and M. Parrinello, “Generalized neural-network representation of high-dimensional potential-energy surfaces,” Phys. Rev. Lett., vol. 98, p. 146401, Apr 2007. [Online]. Available: https://link.aps.org/doi/10.1103/PhysRevLett.98.146401
  15. J. Behler, “First principles neural network potentials for reactive simulations of large molecular and condensed systems,” Angewandte Chemie International Edition, vol. 56, no. 42, pp. 12 828–12 840, 2017.
  16. K. Yao, J. E. Herr, S. N. Brown, and J. Parkhill, “Intrinsic bond energies from a bonds-in-molecules neural network,” The Journal of Physical Chemistry Letters, vol. 8, no. 12, pp. 2689–2694, 2017, pMID: 28573865. [Online]. Available: https://doi.org/10.1021/acs.jpclett.7b01072
  17. J. S. Smith, O. Isayev, and A. E. Roitberg, “Ani-1: an extensible neural network potential with dft accuracy at force field computational cost,” Chemical science, vol. 8, no. 4, pp. 3192–3203, 2017.
  18. S. Desai, S. T. Reeve, and J. F. Belak, “Implementing a neural network interatomic model with performance portability for emerging exascale architectures,” Computer Physics Communications, vol. 270, p. 108156, 2022. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S001046552100268X
  19. Y.-P. Huang, Y. Xia, L. Yang, J. Wei, Y. I. Yang, and Y. Q. Gao, “Sponge: A gpu-accelerated molecular dynamics package with enhanced sampling and ai-driven algorithms,” Chinese Journal of Chemistry, vol. 40, no. 1, pp. 160–168, 2022. [Online]. Available: https://onlinelibrary.wiley.com/doi/abs/10.1002/cjoc.202100456
  20. H. Wang, L. Zhang, J. Han, and E. Weinan, “Deepmd-kit: A deep learning package for many-body potential energy representation and molecular dynamics,” Computer Physics Communications, vol. 228, pp. 178–184, 2018.
  21. K. Schütt, P.-J. Kindermans, H. E. Sauceda Felix, S. Chmiela, A. Tkatchenko, and K.-R. Müller, “Schnet: A continuous-filter convolutional neural network for modeling quantum interactions,” Advances in neural information processing systems, vol. 30, 2017.
  22. R. Drautz, “Atomic cluster expansion for accurate and transferable interatomic potentials,” Phys. Rev. B, vol. 99, p. 014104, Jan 2019. [Online]. Available: https://link.aps.org/doi/10.1103/PhysRevB.99.014104
  23. Y. Lysogorskiy, C. v. d. Oord, A. Bochkarev, S. Menon, M. Rinaldi, T. Hammerschmidt, M. Mrovec, A. Thompson, G. Csányi, C. Ortner, and R. Drautz, “Performant implementation of the atomic cluster expansion (pace) and application to copper and silicon,” npj Computational Materials, vol. 7, no. 1, p. 97, Jun 2021. [Online]. Available: https://doi.org/10.1038/s41524-021-00559-9
  24. S. Batzner, A. Musaelian, L. Sun, M. Geiger, J. P. Mailoa, M. Kornbluth, N. Molinari, T. E. Smidt, and B. Kozinsky, “E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials,” Nature Communications, vol. 13, no. 1, p. 2453, May 2022. [Online]. Available: https://doi.org/10.1038/s41467-022-29939-5
  25. J. Gasteiger, J. Groß, and S. Günnemann, “Directional message passing for molecular graphs,” in International Conference on Learning Representations (ICLR), 2020.
  26. J. Gasteiger, S. Giri, J. T. Margraf, and S. Günnemann, “Fast and uncertainty-aware directional message passing for non-equilibrium molecules,” in Machine Learning for Molecules Workshop, NeurIPS, 2020.
  27. O. T. Unke, S. Chmiela, M. Gastegger, K. T. Schütt, H. E. Sauceda, and K.-R. Müller, “Spookynet: Learning force fields with electronic degrees of freedom and nonlocal effects,” Nature Communications, vol. 12, no. 1, p. 7273, Dec 2021. [Online]. Available: https://doi.org/10.1038/s41467-021-27504-0
  28. A. Musaelian, S. Batzner, A. Johansson, L. Sun, C. J. Owen, M. Kornbluth, and B. Kozinsky, “Learning local equivariant representations for large-scale atomistic dynamics,” Nature Communications, vol. 14, no. 1, p. 579, 2023.
  29. D. Zhang, X. Liu, X. Zhang, C. Zhang, C. Cai, H. Bi, Y. Du, X. Qin, J. Huang, B. Li et al., “Dpa-2: Towards a universal large atomic model for molecular and material simulation,” arXiv preprint arXiv:2312.15492, 2023.
  30. A. Musaelian, A. Johansson, S. Batzner, and B. Kozinsky, “Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size,” arXiv preprint arXiv:2304.10061, 2023.
  31. K. Nguyen-Cong, J. T. Willman, S. G. Moore, A. B. Belonoshko, R. Gayatri, E. Weinberg, M. A. Wood, A. P. Thompson, and I. I. Oleynik, “Billion atom molecular dynamics simulations of carbon at extreme conditions and experimental time and length scales,” in Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2021, pp. 1–12.
  32. Z. Guo, D. Lu, Y. Yan, S. Hu, R. Liu, G. Tan, N. Sun, W. Jiang, L. Liu, Y. Chen et al., “Extending the limit of molecular dynamics with ab initio accuracy to 10 billion atoms,” in Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, 2022, pp. 205–218.
  33. W. Jia, H. Wang, M. Chen, D. Lu, L. Lin, R. Car, E. Weinan, and L. Zhang, “Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning,” in SC20: International conference for high performance computing, networking, storage and analysis.   IEEE, 2020, pp. 1–14.
  34. J. Zeng, D. Zhang, D. Lu, P. Mo, Z. Li, Y. Chen, M. Rynik, L. Huang, Z. Li, S. Shi et al., “Deepmd-kit v2: A software package for deep potential models,” The Journal of Chemical Physics, vol. 159, no. 5, 2023.
  35. M. Galib and D. T. Limmer, “Elucidating the mechanism of reactive uptake of n _⁢2_2\_2_ 2 o _⁢5_5\_5_ 5 in aqueous aerosol,” arXiv preprint arXiv:2005.10134, 2020.
  36. S. Du, X. You, H. Yang, J. Shang, Z. Xiao, Z. Wu, Z. Luan, and D. Qian, “Efficient deep molecular dynamic model training on heterogeneous system,” in 2023 IEEE 29th International Conference on Parallel and Distributed Systems (ICPADS).   IEEE, 2023, pp. 1869–1876.
  37. A. Singraber, T. Morawietz, J. Behler, and C. Dellago, “Parallel multistream training of high-dimensional neural network potentials,” Journal of chemical theory and computation, vol. 15, no. 5, pp. 3075–3092, 2019.
  38. X. Duan, P. Gao, M. Zhang, T. Zhang, H. Meng, Y. Li, B. Schmidt, H. Fu, L. Gan, W. Xue et al., “Cell-list based molecular dynamics on many-core processors: a case study on sunway taihulight supercomputer,” in SC20: International Conference for High Performance Computing, Networking, Storage and Analysis.   IEEE, 2020, pp. 1–12.
  39. J. Li, T. Zhao, Z. Guo, S. Shi, L. Liu, G. Tan, W. Jia, G. Yuan, and Z. Wang, “Enhance the strong scaling of lammps on fugaku,” in Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 2023, pp. 1–13.
  40. TOP500.org, “Supercomputer fugaku - supercomputer fugaku, a64fx 48c 2.2ghz, tofu interconnect d,” https://www.top500.org/system/179807/, 2023.
  41. D. Lu, W. Jiang, Y. Chen, L. Zhang, W. Jia, H. Wang, and M. Chen, “Dp compress: A model compression scheme for generating efficient deep potential models,” Journal of chemical theory and computation, vol. 18, no. 9, pp. 5559–5567, 2022.
  42. H.-Y. Ko, L. Zhang, B. Santra, H. Wang, W. E, R. A. DiStasio Jr, and R. Car, “Isotope effects in liquid water via deep potential molecular dynamics,” Molecular Physics, vol. 117, no. 22, pp. 3269–3281, 2019.
  43. R. A. DiStasio, B. Santra, Z. Li, X. Wu, and R. Car, “The individual and collective effects of exact exchange and dispersion interactions on the ab initio structure of liquid water,” The Journal of chemical physics, vol. 141, no. 8, 2014.
  44. M. Chen, H.-Y. Ko, R. C. Remsing, M. F. Calegari Andrade, B. Santra, Z. Sun, A. Selloni, R. Car, M. L. Klein, J. P. Perdew et al., “Ab initio theory and modeling of water,” Proceedings of the National Academy of Sciences, vol. 114, no. 41, pp. 10 846–10 851, 2017.
  45. L. Zhang, J. Han, H. Wang, R. Car, and E. Weinan, “Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics,” Physical review letters, vol. 120, no. 14, p. 143001, 2018.
  46. K. J. Bowers, E. Chow, H. Xu, R. O. Dror, M. P. Eastwood, B. A. Gregersen, J. L. Klepeis, I. Kolossvary, M. A. Moraes, F. D. Sacerdoti et al., “Scalable algorithms for molecular dynamics simulations on commodity clusters,” in Proceedings of the 2006 ACM/IEEE Conference on Supercomputing, 2006, pp. 84–es.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets