Fast parallel algorithms for short-range molecular dynamics

Published on Mar 1, 1995in Journal of Computational Physics3.553
路 DOI :10.1006/JCPH.1995.1039
Steven J. Plimpton43
Estimated H-index: 43
(SNL: Sandia National Laboratories)
Sources
Abstract
Abstract Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently鈥攖hose with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.
Download
馃摉 Papers frequently viewed together
References51
Newest
Efficient use of a distributed memory parallel computer requires that the computational load be balanced across processors in a way that minimizes interprocessor communication. A new domain mapping algorithm is presented that extends recent work in which ideas from spectral graph theory have been applied to this problem. The generalization of spectral graph bisection involves a novel use of multiple eigenvectors to allow for division of a computation into four or eight parts at each stage of a r...
Source
Dense linear systems of equations are quite common in science and engineering, arising in boundary element methods, least squares problems, and other settings. Massively parallel computers will be necessary to solve the large systems required by scientists and engineers, and scalable parallel algorithms for the linear algebra applications must be devised for these machines. A critical step in these algorithms is the mapping of matrix elements to processors. In this paper, the use of the torus-wr...
Source
#1A.B. MaccabeH-Index: 2
#2R. RiesenH-Index: 2
Last. K.S. McCurleyH-Index: 1
view all 4 authors...
SUNMOS is an acronym for Sandia/UNM Operating System. It was originally developed for the nCUBE-2 MIMD supercomputer between January and December of 1991. Between April and August of 1993, SUNMOS was ported to the Intel Paragon. This document provides a quick overview of how to compile and run jobs using the SUNMOS environment on the Paragon. The primary goal of SUNMOS is to provide high performance message passing and process support an example of its capabilities, SUNMOS Release 1.4 occupies a...
Source
#1Mike BarnettH-Index: 25
#2Lance ShulerH-Index: 3
Last. Jerrell WattsH-Index: 11
view all 6 authors...
We outline a unified approach for building a library of collective communication operations that performs well on a cross-section of problems encountered in real applications. The target architecture is a two-dimensional mesh with worm-hole routing, but the techniques also apply to higher dimensional meshes and hypercubes. We stress a general approach, addressing the need for implementations that perform well for various sized vectors and grid dimensions, including non-power-of-two grids. This r...
Source
#1D. M. Beazley (LANL: Los Alamos National Laboratory)H-Index: 5
#2Peter S. Lomdahl (LANL: Los Alamos National Laboratory)H-Index: 47
Last. Pablo Tamayo (LANL: Los Alamos National Laboratory)H-Index: 6
view all 4 authors...
Presents several techniques that we have used to optimize the performance of a message-passing C code for molecular dynamics on the CM-5. We describe our use of the CM-5 vector units and a parallel memory caching scheme that we have developed to speed up the code by more than 50%. A modification that decreases communication time by 35% is also presented, along with a discussion of how we have been able to take advantage of the CM-5 hardware without significantly compromising code portability. We...
Source
#1J. G. LewisH-Index: 2
#2R.A. van de Geijn (University of Texas at Austin)H-Index: 13
The critical bottlenecks in the implementation of the conjugate gradient algorithm on distributed memory computers are the communication requirements of the sparse matrix-vector multiply and of the vector recurrences. The data distribution and communication patterns of five general implementations whose realizations demonstrate that the cost of communication can be overcome to a much larger extent than is often assumed are described. The results also apply to more general settings for matrix-vec...
Source
#1Aiichiro Nakano (LSU: Louisiana State University)H-Index: 62
#2Priya Vashishta (LSU: Louisiana State University)H-Index: 72
Last. Rajiv K. Kalia (LSU: Louisiana State University)H-Index: 70
view all 3 authors...
Abstract For particles interacting via two- and three-body potentials, a domain-decomposition algorithm is used to implement molecular dynamics (MD) on distributed memory MIMD (multiple-instruction multiple-data) computers. The algorithm employs the linked-cell-list method and separable three-body force calculation. The force calculation is accelerated by the multiple-time-step (MTS) method. For a 1.54 million particle SiO 2 system, the MD program runs at a speed of 660 time steps per hour (1100...
Source
Algorithms for the N-body problem are compared and contrasted, particularly those where N is in the range for which direct methods outperform approximation methods. With fewer bodies than processors, the so-called 鈥渞eplicated orrery鈥 on a three-dimensional grid has been used successfully on the Connection Machine CM-2 architecture. With more bodies, the 鈥渞otated and translated Gray codes鈥 is an ideal direct algorithm for machines such as the CM-2 in that it takes optimal advantage of the communi...
Source
This paper describes an implementation of a parallel molecular dynamics algorithm on the CM2 Connection Machine that is designed for large-scale simulations. The method employs a cell subdivision of the simulation region, and is partly based on the layer approach developed for vector processing. All communication is between adjacent processing elements, eliminating the need for global communication. Performance measurements were made with systems containing over 106 particles.
Source
#1Soumya S. Patnaik (UVA: University of Virginia)H-Index: 19
#2Ruth PachterH-Index: 39
Last. W. Wade AdamsH-Index: 45
view all 4 authors...
We have used molecular dynamics (MD) to study the room temperature bulk phase behavior of a cyclic siloxane with a pentamethylcyclosiloxane core and biphenyl-4-allyloxybenzoate Mesogens (BCS). This Material exhibits thermotropic liquid crystalline behavior above 120 掳C. Bonded and non-bonded interactions were considered and a Molecular Mechanics force field was used to model the structural anisotropy of the siloxane Molecules. Molecular clusters with and without periodic boundary conditions (pbc...
Source
Cited By25997
Newest
Last. Liu Fukang (NPU: Northwestern Polytechnical University)H-Index: 5
view all 6 authors...
Abstract null null The strength of polycrystalline metals increases with decreasing grain size, following the classical Hall-Petch relationship. However, this relationship fails when softening occurs at very small grain sizes (typically less than 10 to 20 nm), which limits the development of ultrahigh-strength materials. In this work, using columnar-grained nanocrystalline Cu-Ag 鈥榮amples鈥, molecular dynamics simulations were performed to investigate the softening mechanism and explore the streng...
Source
#1Jiayong ZhangH-Index: 6
#2Hongwu ZhangH-Index: 7
Last. Jian LuH-Index: 79
view all 7 authors...
Source
#1Yeonju Oh (SNU: Seoul National University)H-Index: 4
#2Won-Seok Ko (UOU: University of Ulsan)H-Index: 15
Last. Heung Nam Han (SNU: Seoul National University)H-Index: 41
view all 6 authors...
Abstract null null Tungsten as a material exhibits broad and increasingly important applications; however, the characterization of its ductile-to-brittle transition (BDT) is currently limited to large-scale scenarios and destructive testing. In this study, we overcome this challenge by implementing small-scale techniques to provide a comprehensive understanding of the BDT behavior of pure tungsten. In order to predict the failure mode at various temperature ranges, the practical fracture analysi...
Source
#1William D. Fullmer (DOE: United States Department of Energy)H-Index: 2
#2Roberto Porcu (DOE: United States Department of Energy)H-Index: 2
Last. Ishan Srivastava (LBNL: Lawrence Berkeley National Laboratory)H-Index: 7
view all 5 authors...
Abstract null null The n-body instability is investigated with the soft-sphere discrete element method. The divergence of nearby trajectories is quantified by the dynamical memory time. Using the inverse proportionality between the dynamical memory time and the largest Lyapunov exponent, the soft-sphere discrete element method results are compared to previous hard-sphere molecular dynamics data for the first time. Good agreement is observed at low concentrations and the degree of instability is ...
Source
#1Jiqiang Wang (HIT: Harbin Institute of Technology)H-Index: 7
#2Yanquan Geng (HIT: Harbin Institute of Technology)H-Index: 14
Last. Pengfei Fan (University of Strathclyde)H-Index: 3
view all 6 authors...
Subsurface damage that is caused by mechanical machining is a major impediment to the widespread use of hard鈥揵rittle materials. Ultrasonic vibration-assisted macro- or micromachining could facilitate shallow subsurface damage compared with conventional machining. However, the subsurface damage that was induced by ultrasonic vibration-assisted nanomachining on hard鈥揵rittle silicon crystal has not yet been thoroughly investigated. In this study, we used a tip-based ultrasonic vibration-assisted na...
Source
#1Seungchan ChoH-Index: 14
#2Junghwan KimH-Index: 19
Last. Sang-Kwan LeeH-Index: 16
view all 11 authors...
Abstract null null This study shows that the mechanical strength of the composite of Fe matrix and titanium carbide (TiC) ceramic particles is significantly enhanced with addition of molybdenum (Mo) atoms. TiC reinforced Fe (Fe-0.2C-7Mn) composites with and without Mo were fabricated by a liquid pressing infiltration (LPI) process and the effect of Mo on interfacial properties of TiC鈥揊e composite was investigated using atomic probe tomography (APT) analysis, molecular dynamics (MD) simulations, ...
Source
#1Junwei Che (Xi'an Jiaotong University)H-Index: 7
#2Xiangyang Liu (THU: Tsinghua University)H-Index: 7
Last. Shengli Zhang (Xi'an Jiaotong University)H-Index: 29
view all 7 authors...
Abstract null null Thermally grown oxides (TGOs) at the ceramic top-coat/metallic bond-coat interface are a pressing challenge in advanced thermal barrier coating (TBC) systems as they can affect the performance and service lifetime of TBCs. Thus, developing novel TBC materials with ultralow oxygen ion diffusivity is very urgent. In this study, we reported the diffusive properties of oxygen ions in a novel pyrochlore-type La2(Zr0.7Ce0.3)2O7 (LZ7C3) material. The measured ionic conductivity and a...
Source
#1Wu-Rong Jian (UCSB: University of California, Santa Barbara)H-Index: 12
#2Zhuocheng Xie (SCUT: South China University of Technology)H-Index: 2
Last. Irene J. Beyerlein (UCSB: University of California, Santa Barbara)H-Index: 82
view all 5 authors...
Abstract null null We perform molecular dynamics simulations to investigate shock-induced amorphization in CoCrNi, a medium entropy alloy (MEA) and its mean-field variant without lattice distortion. We show that a critical velocity exists above which amorphization occurs. At a low shock velocity of 800聽m/s, dislocation slip and twins dominate and amorphization does not happen, but as the shock velocity increases, the deformation mechanism transitions from slip and twinning to solid-state amorphi...
Source
#1Akhilesh Gandhi (A&M: Texas A&M University)H-Index: 1
#2M. M. Faruque Hasan (A&M: Texas A&M University)H-Index: 25
Porous crystalline materials, such as zeolites and metal-organic frameworks (MOFs), have shown great promises with superior separation, catalysis and upgrading performances in many areas of energy, the environment and health. However, the discovery of new zeolites and MOFs with desired properties is a complex process that often involves trial-and-error experimental/computational approaches. Computational discovery of new materials often involves learning and optimizing more than one objective su...
Source
#1Xiao-Feng Yang (ECUST: East China University of Science and Technology)
#2Yongzhi Xi (ECUST: East China University of Science and Technology)
Last. Shan-Tung Tu (ECUST: East China University of Science and Technology)H-Index: 51
view all 6 authors...
Abstract null null The strengthening effect of chemical short-range order (SRO) structure in CoCrNi medium-entropy alloy (MEA) was investigated using molecular dynamics (MD) simulations of nanoindentation. The quantitative correlation between SRO parameters and mechanical properties was established. Results show that the strength and hardness of CoCrNi MEA increase with increasing chemical SRO parameters and reach a stable value with steady SRO structure. Compared with random solid solution (RSS...
Source
This website uses cookies.
We use cookies to improve your online experience. By continuing to use our website we assume you agree to the placement of these cookies.
To learn more, you can find in our Privacy Policy.