The parallel replica dynamics method – Coming of age

作者: Danny Perez , Blas P. Uberuaga , Arthur F. Voter

DOI: 10.1016/J.COMMATSCI.2014.12.011

关键词:

摘要: Abstract Molecular dynamics (MD) – the numerical integration of atomistic equations motion is a workhorse computational materials science. Indeed, MD can in principle be used to obtain any thermodynamic or kinetic quantity, without introducing approximation assumptions beyond adequacy interaction potential. It therefore an extremely powerful and flexible tool study with spatio-temporal resolution. These enviable qualities however come at steep price, limiting system sizes simulation times that achieved practice. While size limitation efficiently addressed massively parallel implementations based on spatial decomposition strategies, allowing for trillions atoms, same approach usually cannot extend timescales much microseconds. In this article, we discuss alternative, parallel-in-time, strategy Parallel Replica Dynamics (ParRep) method aims addressing timescale systems evolve through rare state-to-state transitions. We review formal underpinnings method, including recent developments showing it provide arbitrarily accurate results definition states. When adequate states available, ParRep simulate trajectories speedup approaching number replicas used. demonstrate usefulness by presenting different examples simulations where access long was essential physical regime interest practical considerations must carry out these simulations. Sixteen years after its introduction, new understanding generality ever increasing availability processing, coming age.

参考文章(67)
Fedwa El-Mellouhi, Normand Mousseau, Laurent J. Lewis, Kinetic activation-relaxation technique: An off-lattice self-learning kinetic Monte Carlo algorithm Physical Review B. ,vol. 78, pp. 153202- ,(2008) , 10.1103/PHYSREVB.78.153202
Blas Pedro Uberuaga, Steven J. Stuart, Arthur F. Voter, Parallel replica dynamics for driven systems : Derivation and application to strained nanotubes Physical Review B. ,vol. 75, pp. 014301- ,(2007) , 10.1103/PHYSREVB.75.014301
Cristian Sminchisescu, Bill Triggs, Fast mixing hyperdynamic sampling Image and Vision Computing. ,vol. 24, pp. 279- 289 ,(2006) , 10.1016/J.IMAVIS.2005.07.022
Chun-Li Liu, James B. Adams, Structure and diffusion of clusters on Ni surfaces Surface Science. ,vol. 268, pp. 73- 86 ,(1992) , 10.1016/0039-6028(92)90951-2
Claude Le Bris, Tony Lelievre, Mitchell Luskin, Danny Perez, A mathematical formalization of the parallel replica dynamics Monte Carlo Methods and Applications. ,vol. 18, pp. 119- 146 ,(2012) , 10.1515/MCMA-2012-0003
Yunsic Shim, Jacques G. Amar, Adaptive temperature-accelerated dynamics The Journal of Chemical Physics. ,vol. 134, pp. 054127- 054127 ,(2011) , 10.1063/1.3541823
Philipp Metzner, Marcus Weber, Christof Schütte, Observation uncertainty in reversible Markov chains. Physical Review E. ,vol. 82, pp. 031114- ,(2010) , 10.1103/PHYSREVE.82.031114
Chun-Yaung Lu, Dmitrii E. Makarov, Graeme Henkelman, Communication: κ-dynamics--an exact method for accelerating rare event classical molecular dynamics. Journal of Chemical Physics. ,vol. 133, pp. 201101- 201101 ,(2010) , 10.1063/1.3514030
Brian Puchala, Michael L. Falk, Krishna Garikipati, An energy basin finding algorithm for kinetic Monte Carlo acceleration Journal of Chemical Physics. ,vol. 132, pp. 134104- 134104 ,(2010) , 10.1063/1.3369627
J. C. Hamilton, M. S. Daw, S. M. Foiles, Dislocation Mechanism for Island Diffusion on fcc (111) Surfaces. Physical Review Letters. ,vol. 74, pp. 2760- 2763 ,(1995) , 10.1103/PHYSREVLETT.74.2760