Accurate and efficient molecular dynamics based on machine learning and non von Neumann architecture
Source abstract
Abstract
Abstract Force field-based classical molecular dynamics (CMD) is efficient but its potential energy surface (PES) prediction error can be very large. Density functional theory (DFT)-based ab-initio molecular dynamics (AIMD) is accurate but computational cost limits its applications to small systems. Here, we propose a molecular dynamics (MD) methodology which can simultaneously achieve both AIMD-level high accuracy and CMD-level high efficiency. The high accuracy is achieved by exploiting deep neural network (DNN)’s arbitrarily-high precision to fit PES. The high efficiency is achieved by deploying multiplication-less DNN on a carefully-optimized special-purpose non von Neumann (NvN) computer to mitigate the performance-limiting data shuttling (i.e., ‘memory wall bottleneck’). By testing on different molecules and bulk systems, we show that the proposed MD methodology is generally-applicable to various MD tasks. The proposed MD methodology has been deployed on an in-house computing server based on reconfigurable field programmable gate array (FPGA), which is freely available athttp://nvnmd.picp.vip.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Machine Learning in Materials Science · Advanced Memory and Neural Computing · Ferroelectric and Negative Capacitance Devices
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (100+)
Show source-provided references
1. Bapst, V. et al. Unveiling the predictive power of static structure in glassy systems. Nat. Phys. 16, 448–454 (2020).10.1038/s41567-020-0842-8 ↗
2. Schott, S. et al. Polaron spin dynamics in high-mobility polymeric semiconductors. Nat. Phys. 15, 814–822 (2019).10.1038/s41567-019-0538-0 ↗
3. Galib, M. & Limmer, D. T. Reactive uptake of N2O5 by atmospheric aerosol is dominated by interfacial processes. Science 371, 921–925 (2021).10.1126/science.abd7716 ↗
4. Widmer, D. R. & Schwartz, B. J. Solvents can control solute molecular identity. Nat. Chem. 10, 910–916 (2018).10.1038/s41557-018-0066-z ↗
5. Karplus, M. & Petsko, G. A. Molecular dynamics simulations in biology. Nature 347, 631–639 (1990).10.1038/347631a0 ↗
6. Chen, S. et al. Simultaneously enhancing the ultimate strength and ductility of high-entropy alloys via short-range ordering. Nat. Commun. 12, 4953 (2021).10.1038/s41467-021-25264-5 ↗
7. Ding, W. et al. Prediction of intrinsic two-dimensional ferroelectrics in In2Se3 and other III2-VI3 van der Waals materials. Nat. Commun. 8, 14956 (2017).10.1038/ncomms14956 ↗
8. Wang, Y. et al. Dynamic deformability of individual PbSe nanocrystals during superlattice phase transitions. Sci. Adv. 5, eaaw5623 (2019).10.1126/sciadv.aaw5623 ↗
9. Lehtinen, O., Kurasch, S., Krasheninnikov, A. V. & Kaiser, U. Atomic scale study of the life cycle of a dislocation in graphene from birth to annihilation. Nat. Commun. 4, 2098 (2013).10.1038/ncomms3098 ↗
10. Lu, S. et al. Activation pathway of a G protein-coupled receptor uncovers conformational intermediates as targets for allosteric drug design. Nat. Commun. 12, 4721 (2021).10.1038/s41467-021-25020-9 ↗
11. Zhao, Y. et al. Augmenting drug–carrier compatibility improves tumour nanotherapy efficacy. Nat. Commun. 7, 11221 (2016).10.1038/ncomms11221 ↗
12. Laio, A., Bernard, S., Chiarotti, G. L., Scandolo, S. & Tosatti, E. Physics of iron at Earth’s core conditions. Science 287, 1027–1030 (2000).10.1126/science.287.5455.1027 ↗
13. Steinle-Neumann, G., Stixrude, L., Cohen, R. E. & Gülseren, O. Elasticity of iron at the temperature of the Earth’s inner core. Nature 413, 57–60 (2001).10.1038/35092536 ↗
14. Hughes, M. A. et al. n-type chalcogenides by ion implantation. Nat. Commun. 5, 5346 (2014).10.1038/ncomms6346 ↗
15. Wang, X.-P. et al. Time-dependent density-functional theory molecular-dynamics study on amorphization of Sc-Sb-Te alloy under optical excitation. npj Comput. Mater. 6, 31 (2020).10.1038/s41524-020-0303-z ↗
16. Kohn, W. & Sham, L. J. Self-consistent equations including exchange and correlation effects. Phys. Rev. 140, A1133–A1138 (1965).10.1103/PhysRev.140.A1133 ↗
17. Car, R. & Parrinello, M. Unified approach for molecular dynamics and density-functional theory. Phys. Rev. Lett. 55, 2471–2474 (1985).10.1103/PhysRevLett.55.2471 ↗
18. Alavi, S. Ab initio molecular dynamics basic theory and advanced methods. By Dominik Marx and Jürg Hutter. Angew. Chem. Int. Ed. 48, 9404–9405 (2009).10.1002/anie.200904748 ↗
19. Jorgensen, W. L., Maxwell, D. S. & Tirado-Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 118, 11225–11236 (1996).10.1021/ja9621760 ↗
20. Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A. & Case, D. A. Development and testing of a general Amber force field. J. Comput. Chem. 25, 1157–1174 (2004).10.1002/jcc.20035 ↗
21. Vanommeslaeghe, K. et al. CHARMM general force field: a force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields. J. Comput. Chem. 31, 671–690 (2010).10.1002/jcc.21367 ↗
22. Shaw, D. E. et al. Anton, a special-purpose machine for molecular dynamics simulation. Commun. ACM 51, 91–97 (2008).10.1145/1364782.1364802 ↗
23. Shaw, D. E. et al. Anton 2: Raising the Bar for Performance and Programmability in a Special-Purpose Molecular Dynamics Supercomputer. in SC14: International Conference for High Performance Computing, Networking, Storage and Analysis 2015-January, 41–53 (IEEE, 2014).
24. Shaw, D. E. et al. Anton 3: twenty microseconds of molecular dynamics simulation before lunch. in Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis 1–11 (ACM, 2021). https://doi.org/10.1145/3458817.3487397.10.1145/3458817.3487397 ↗
25. Behler, J. & Parrinello, M. Generalized neural-network representation of high-dimensional potential-energy surfaces. Phys. Rev. Lett. 98, 146401 (2007).10.1103/PhysRevLett.98.146401 ↗
26. Wang, H., Zhang, L., Han, J. & E, W. DeePMD-kit: a deep learning package for many-body potential energy representation and molecular dynamics. Comput. Phys. Commun. 228, 178–184 (2018).10.1016/j.cpc.2018.03.016 ↗
27. Zhang, L., Han, J., Wang, H., Car, R. & E, W. Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics. Phys. Rev. Lett. 120, 143001 (2018).10.1103/PhysRevLett.120.143001 ↗
28. Zhang, L., Lin, D.-Y., Wang, H., Car, R. & E, W. Active learning of uniformly accurate interatomic potentials for materials simulation. Phys. Rev. Mater. 3, 023804 (2019).10.1103/PhysRevMaterials.3.023804 ↗
29. Zhang, Y. et al. DP-GEN: a concurrent learning platform for the generation of reliable deep learning based potential energy models. Comput. Phys. Commun. 253, 107206 (2020).10.1016/j.cpc.2020.107206 ↗
30. Zhang, L. et al. End-to-end Symmetry Preserving Inter-atomic Potential Energy Model for Finite and Extended Systems. Adv. Neural Inf. Process. Syst. 2018-December, 4436–4446 (2018).
31. Jia, W. et al. 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 1–14 (IEEE, 2020). https://doi.org/10.1109/SC41405.2020.00009.10.1109/SC41405.2020.00009 ↗
32. LAMMPS Benchmarks. Available at: https://www.lammps.org/bench.html.
33. Wulf, W. A. & McKee, S. A. Hitting the memory wall. ACM SIGARCH Comput. Archit. N. 23, 20–24 (1995).10.1145/216585.216588 ↗
34. Horowitz, M. 1.1 Computing’s energy problem (and what we can do about it). in 2014 IEEE International Solid-State Circuits Conference Digest of Technical Papers (ISSCC) 57, 10–14 (IEEE, 2014).10.1109/ISSCC.2014.6757323 ↗
35. Ielmini, D. & Wong, H. S. P. In-memory computing with resistive switching devices. Nat. Electron. 1, 333–343 (2018).10.1038/s41928-018-0092-2 ↗
36. Stegailov, V., Smirnov, G. & Vecher, V. VASP hits the memory wall: processors efficiency comparison. Concurr. Comput. Pract. Exp. 31, e5136 (2019).10.1002/cpe.5136 ↗
37. John von Neumann. First Draft of a Report on the EDVAC. (1945).10.5479/sil.538961.39088011475779 ↗
38. Electronic Numerical Integrator and Computer (ENIAC). Available at: https://en.wikipedia.org/wiki/ENIAC.
39. Beyond von Neumann. Nat. Nanotechnol. 15, 507–507 (2020).10.1038/s41565-020-0738-x ↗
40. Taiji, M. et al. Protein Explorer: A Petaflops Special-Purpose Computer System for Molecular Dynamics Simulations. in Proceedings of the 2003 ACM/IEEE conference on Supercomputing - SC ’03 15 (ACM Press, 2003). https://doi.org/10.1145/1048935.1050166.10.1145/1048935.1050166 ↗
41. Harvey, M. J., Giupponi, G. & De Fabritiis, G. ACEMD: Accelerating biomolecular dynamics in the microsecond time scale. J. Chem. Theory Comput. 5, 1632–1639 (2009).10.1021/ct9000685 ↗
42. Deringer, V. L. & Csányi, G. Machine learning based interatomic potential for amorphous carbon. Phys. Rev. B 95, 094203 (2017).10.1103/PhysRevB.95.094203 ↗
43. Rowe, P., Csányi, G., Alfè, D. & Michaelides, A. Development of a machine learning potential for graphene. Phys. Rev. B 97, 054303 (2018).10.1103/PhysRevB.97.054303 ↗
44. Zeng, J., Cao, L., Xu, M., Zhu, T. & Zhang, J. Z. H. Complex reaction processes in combustion unraveled by neural network-based molecular dynamics simulation. Nat. Commun. 11, 5713 (2020).10.1038/s41467-020-19497-z ↗
45. Li, R., Lee, E. & Luo, T. A unified deep neural network potential capable of predicting thermal conductivity of silicon in different phases. Mater. Today Phys. 12, 100181 (2020).10.1016/j.mtphys.2020.100181 ↗
46. Rowe, P., Deringer, V. L., Gasparotto, P., Csányi, G. & Michaelides, A. An accurate and transferable machine learning potential for carbon. J. Chem. Phys. 153, 034702 (2020).10.1063/5.0005084 ↗
47. Bettini, J. et al. Experimental realization of suspended atomic chains composed of different atomic species. Nat. Nanotechnol. 1, 182–185 (2006).10.1038/nnano.2006.132 ↗
48. Wu, B. et al. Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutions. in 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition 9127–9135 (IEEE, 2018). https://doi.org/10.1109/CVPR.2018.0095110.1109/CVPR.2018.00951 ↗
49. Chen, H. et al. AdderNet: Do We Really Need Multiplications in Deep Learning? in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 1465–1474 (IEEE, 2020). https://doi.org/10.1109/CVPR42600.2020.0015410.1109/CVPR42600.2020.00154 ↗
50. Ahn, J., Yoo, S., Mutlu, O. & Choi, K. PIM-enabled instructions. in Proceedings of the 42nd Annual International Symposium on Computer Architecture 43, 336–348 (ACM, 2015).10.1145/2872887.2750385 ↗
51. Mutlu, O., Ghose, S., Gómez-Luna, J. & Ausavarungnirun, R. Processing data where it makes sense: Enabling in-memory computation. Microprocess. Microsyst. 67, 28–41 (2019).10.1016/j.micpro.2019.01.009 ↗
52. Liu, J. & Mo, P. The server website of NVNMD. (2021). Available at: http://nvnmd.picp.vip/.
53. Liu, J. & Mo, P. The training and testing code for NVNMD. (2021). Available at: https://github.com/LiuGroupHNU/nvnmd.
54. Abadi, M. et al. TensorFlow: A system for large-scale machine learning. Proc. 12th USENIX Symp. Oper. Syst. Des. Implementation, OSDI 2016 265–283 (2016). https://doi.org/10.5555/3026877.302689910.5555/3026877.3026899 ↗
55. Sosso, G. C., Miceli, G., Caravati, S., Behler, J. & Bernasconi, M. Neural network interatomic potential for the phase change material GeTe. Phys. Rev. B 85, 174103 (2012).10.1103/PhysRevB.85.174103 ↗
56. Shi, M., Mo, P. & Liu, J. Deep Neural Network for Accurate and Efficient Atomistic Modeling of Phase Change Memory. IEEE Electron Device Lett. 41, 365–368 (2020).10.1109/LED.2020.2964779 ↗
57. Plimpton, S. Fast parallel algorithms for short-range molecular dynamics. Journal of Computational Physics 117, (1993).10.2172/10176421 ↗
58. Chmiela, S. et al. Machine learning of accurate energy-conserving molecular force fields. Sci. Adv. 3, e1603015 (2017).10.1126/sciadv.1603015 ↗
59. Chmiela, S., Sauceda, H. E., Müller, K.-R. & Tkatchenko, A. Towards exact molecular dynamics simulations with machine-learned force fields. Nat. Commun. 9, 3887 (2018).10.1038/s41467-018-06169-2 ↗
60. Christensen, A. S. & von Lilienfeld, O. A. On the role of gradients for machine learning of molecular energies and forces. Mach. Learn. Sci. Technol. 1, 045018 (2020).10.1088/2632-2153/abba6f ↗
61. Shi, M., Li, J., Tao, M., Zhang, X. & Liu, J. Artificial intelligence model for efficient simulation of monatomic phase change material antimony. Mater. Sci. Semicond. Process. 136, 106146 (2021).10.1016/j.mssp.2021.106146 ↗
62. Huang, J. et al. Deep potential generation scheme and simulation protocol for the Li10GeP2S12-type superionic conductors. J. Chem. Phys. 154, 094703 (2021).10.1063/5.0041849 ↗
63. Bogojeski, M., Vogt-Maranto, L., Tuckerman, M. E., Müller, K.-R. & Burke, K. Quantum chemical accuracy from density functional approximations via machine learning. Nat. Commun. 11, 5223 (2020).10.1038/s41467-020-19093-1 ↗
64. Narayanan, B., Redfern, P. C., Assary, R. S. & Curtiss, L. A. Accurate quantum chemical energies for 133000 organic molecules. Chem. Sci. 10, 7449–7455 (2019).10.1039/C9SC02834J ↗
65. Morawietz, T. & Artrith, N. Machine learning-accelerated quantum mechanics-based atomistic simulations for industrial applications. J. Comput. -Aided Mol. Des. 35, 557–586 (2021).10.1007/s10822-020-00346-6 ↗
66. Zhang, P., Shen, L. & Yang, W. Solvation Free Energy Calculations with Quantum Mechanics/Molecular Mechanics and Machine Learning Models. J. Phys. Chem. B 123, 901–908 (2019).10.1021/acs.jpcb.8b11905 ↗
67. Lu, C. et al. OPLS4: Improving force field accuracy on challenging regimes of chemical space. J. Chem. Theory Comput. 17, 4291–4300 (2021).10.1021/acs.jctc.1c00302 ↗
68. Soler, J. M. et al. The SIESTA method for ab initio order-N materials simulation. J. Phys. Condens. Matter 14, 2745–2779 (2002).10.1088/0953-8984/14/11/302 ↗
69. Giannozzi, P. et al. QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials. J. Phys. Condens. Matter 21, 395502 (2009).10.1088/0953-8984/21/39/395502 ↗
70. VandeVondele, J. et al. Quickstep: Fast and accurate density functional calculations using a mixed Gaussian and plane waves approach. Comput. Phys. Commun. 167, 103–128 (2005).10.1016/j.cpc.2004.12.014 ↗
71. Ahn, S. Phase Change Memory. (Springer International Publishing, 2018). https://doi.org/10.1007/978-3-319-69053-7.10.1007/978-3-319-69053-7 ↗
72. Kolobov, A. V., Krbal, M., Fons, P., Tominaga, J. & Uruga, T. Distortion-triggered loss of long-range order in solids with bonding energy hierarchy. Nat. Chem. 3, 311–316 (2011).10.1038/nchem.1007 ↗
73. Mo, Y., Ong, S. P. & Ceder, G. First principles study of the Li 10GeP 2S 12 lithium super ionic conductor material. Chem. Mater. 24, 15–17 (2012).10.1021/cm203303y ↗
74. Marcolongo, A., Binninger, T., Zipoli, F. & Laino, T. Simulating Diffusion Properties of Solid‐State Electrolytes via a Neural Network Potential: Performance and Training Scheme. ChemSystemsChem 2, e1900031 (2020).10.1002/syst.201900031 ↗
75. Kamaya, N. et al. A lithium superionic conductor. Nat. Mater. 10, 682–686 (2011).10.1038/nmat3066 ↗
76. NVIDIA Corporation. Nvidia Tesla V100 GPU Volta Architecture. White Paper 53 (2017). Available at: https://images.nvidia.cn/content/volta-architecture/pdf/volta-architecture-whitepaper.pdf.
77. Summit. Available at: https://www.olcf.ornl.gov/olcf-resources/compute-systems/summit/.
78. NVIDIA. NVIDIA V100. Available at: https://www.nvidia.com/en-us/data-center/v100/.
79. Xilinx. UltraScale Architecture and Product Data Sheet: Overview. Xilinx.com 1–46 (2020). Available at: https://www.xilinx.com/support/documentation/data_sheets/ds890-ultrascale-overview.pdf.
80. Xilinx. UltraScale+ FPGAs Product Tables and Product Selection Guide. Xilinx.com 1–11 (2021). Available at: https://www.xilinx.com/support/documentation/selection-guides/ultrascale-plus-fpga-product-selection-guide.pdf.
81. Ic, S. P., Dube, B., Elisabeth, S. & Scansen, D. Apple M1 System-on-Chip. systemplus.fr 1–36 (2020). Available at: https://www.systemplus.fr/wp-content/uploads/2020/12/SP20608-Apple-M1-System-on-Chip-Sample.pdf.
82. Lu, D. et al. 86 PFLOPS Deep Potential Molecular Dynamics simulation of 100 million atoms with ab initio accuracy. Comput. Phys. Commun. 259, 107624 (2021).10.1016/j.cpc.2020.107624 ↗
83. Samir, N. et al. ASIC and FPGA Comparative Study for IoT lightweight hardware security algorithms. J. Circuits, Syst. Comput. 28, (2019).10.1142/S0218126619300095 ↗
84. Schütt, K. T. et al. SchNet: A continuous-filter convolutional neural network for modeling quantum interactions. Adv. Neural Inf. Process. Syst. 2017-Decem, 992–1002 (2017).
85. Klicpera, J., Groß, J. & Günnemann, S. Directional Message Passing for Molecular Graphs. Preprint at http://arxiv.org/abs/2003.03123 (2020).
86. Chmiela, S., Sauceda, H. E., Poltavsky, I., Müller, K. R. & Tkatchenko, A. sGDML: Constructing accurate and data efficient molecular force fields using machine learning. Comput. Phys. Commun. 240, 38–45 (2019).10.1016/j.cpc.2019.02.007 ↗
87. Schütt, K., Unke, O. & Gastegger, M. Equivariant message passing for the prediction of tensorial properties and molecular spectra. in Proceedings of the 38th International Conference on Machine Learning (Vol. 139 eds. Meila, M. & Zhang, T.) 9377–9388 (PMLR, 2021).
88. Unke, O. T. et al. SpookyNet: Learning force fields with electronic degrees of freedom and nonlocal effects. Nat. Commun. 12, 7273 (2021).10.1038/s41467-021-27504-0 ↗
89. Klicpera, J., Becker, F. & Günnemann, S. GemNet: Universal Directional Graph Neural Networks for Molecules. Preprint at http://arxiv.org/abs/2106.08903 (2021).
90. Haghighatlari, M. et al. NewtonNet: A Newtonian message passing network for deep learning of interatomic potentials and forces. Preprint at http://arxiv.org/abs/2108.02913 (2021).
91. Qiao, Z. et al. UNiTE: Unitary N-body Tensor Equivariant Network with Applications to Quantum Chemistry. Preprint at http://arxiv.org/abs/2105.14655 (2021).
92. Batzner, S. et al. E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials. Preprint at http://arxiv.org/abs/2101.03164 (2021).10.21203/rs.3.rs-244137/v1 ↗
93. Kanal, I. Y., Keith, J. A. & Hutchison, G. R. A sobering assessment of small‐molecule force field methods for low energy conformer predictions. Int. J. Quantum Chem. 118, e25512 (2018).10.1002/qua.25512 ↗
94. Zgarbová, M., Otyepka, M., Šponer, J., Hobza, P. & Jurečka, P. Large-scale compensation of errors in pairwise-additive empirical force fields: Comparison of AMBER intermolecular terms with rigorous DFT-SAPT calculations. Phys. Chem. Chem. Phys. 12, 10476–10493 (2010).10.1039/c002656e ↗
95. Demir, H. et al. DFT-based force field development for noble gas adsorption in metal organic frameworks. J. Mater. Chem. A 3, 23539–23548 (2015).10.1039/C5TA06201B ↗
96. Shen, L. & Yang, W. Molecular Dynamics Simulations with Quantum Mechanics/Molecular Mechanics and Adaptive Neural Networks. J. Chem. Theory Comput. 14, 1442–1455 (2018).10.1021/acs.jctc.7b01195 ↗
97. Jinnouchi, R., Karsai, F. & Kresse, G. Making free-energy calculations routine: combining first principles with machine learning. Phys. Rev. B 101, 060201 (2020).10.1103/PhysRevB.101.060201 ↗
98. Han, J., Zhang, L., Car, R. & E, W. Deep potential: a general representation of a many-body potential energy surface. Commun. Comput. Phys. 23, 629–639 (2018).10.4208/cicp.OA-2017-0213 ↗
99. Allen, M. P. & Tildesley, D. J. Computer Simulation of Liquids. 1, (Oxford University Press, 2017).10.1093/oso/9780198803195.001.0001 ↗
100. Parrinello, M. & Rahman, A. Polymorphic transitions in single crystals: a new molecular dynamics method. J. Appl. Phys. 52, 7182–7190 (1981).10.1063/1.328693 ↗