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—those 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.

Fast parallel algorithms for short-range molecular dynamics

다중혈관 관상동맥 환자에서 y-문합을 이용하여 양쪽 내흉동맥만을 사용한 우회술의 조기 성적

Molecular simulation is an extremely useful, but computationally very expensive tool for studies of chemical and biomolecular systems Here, we present a new implementation of our molecular simulation toolkit GROMACS which now both achieves extremely high performance on single processors from algorithmic optimizations and hand-coded routines and simultaneously scales very well on parallel machines The code encompasses a minimal-communication domain decomposition algorithm, full dynamic load balancing, a state-of-the-art parallel constraint solver, and efficient virtual site algorithms that allow removal of hydrogen atom degrees of freedom to enable integration time steps up to 5 fs for atomistic simulations also in parallel To improve the scaling properties of the common particle mesh Ewald electrostatics algorithms, we have in addition used a Multiple-Program, Multiple-Data approach, with separate node domains responsible for direct and reciprocal space interactions Not only does this combination of a

/pdf/gromacs-4-algorithms-for-highly-efficient-load-balanced-and-16xnd83poh.pdf

GROMACS 4: Algorithms for highly efficient, load-balanced, and scalable molecular simulation

/pdf/scalable-molecular-dynamics-with-namd-3j7xyc70g4.pdf

Scalable Molecular Dynamics with NAMD

http://kth.diva-portal.org/smash/get/diva2:1303357/FULLTEXT01

GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers

Desmond is a molecular dynamics (MD) code designed for high performance on large commodity Linux/Unix clusters. This report documents the performance that Desmond achieved on new hardware in April 2008. For the DHFR system (23,558 atoms) with production simulation parameters, Desmond’s top simulation rate is 471 ns/day on 1024 cores of an InfiniBand cluster. For the ApoA1 system (92,224 atoms), Desmond’s top simulation rate is 289 ns/day on 4096 cores of the same cluster.

/pdf/desmond-performance-on-a-cluster-of-multicore-processors-8exz3ad2so.pdf

Desmond Performance on a Cluster of Multicore Processors

Nonlinear physics governing the kinetic behavior of stimulated Raman scattering (SRS) in multi-speckled laser beams has been identified in the trapping regime over a wide range of kλD values (here k is the wave number of the electron plasma waves and λD is the Debye length) in homogeneous and inhomogeneous plasmas. Hot electrons from intense speckles, both forward and side-loss hot electrons produced during SRS daughter electron plasma wave bowing and filamentation, seed and enhance the growth of SRS in neighboring speckles by reducing Landau damping. Trapping-enhanced speckle interaction through transport of hot electrons, backscatter, and sidescatter SRS light waves enable the system of speckles to self-organize and exhibit coherent, sub-ps SRS bursts with more than 100% instantaneous reflectivity, resulting in an SRS transverse coherence width much larger than a speckle width and a SRS spectrum that peaks outside the incident laser cone. SRS reflectivity is found to saturate above a threshold laser intensity at a level of reflectivity that depends on kλD: higher kλD leads to lower SRS and the reflectivity scales as ∼(kλD)−4. As kλD and Landau damping increase, speckle interaction via sidescattered light and side-loss hot electrons decreases and the occurrence of self-organized events becomes infrequent, leading to the reduction of time-averaged SRS reflectivity. It is found that the inclusion of a moderately strong magnetic field in the laser direction can effectively control SRS by suppressing transverse speckle interaction via hot electron transport.

Self-organized coherent bursts of stimulated Raman scattering and speckle interaction in multi-speckled laser beams

We present large scale 3D particle-in-cell (PIC) simulations to examine particle energization in magnetic reconnection of relativistic electron-positron (pair) plasmas. The initial configuration is set up as a relativistic Harris equilibrium without a guide field. These simulations are large enough to accommodate a sufficient number of tearing and kink modes. Contrary to the non-relativistic limit, the linear tearing instability is faster than the linear kink instability, at least in our specific parameters. We find that the magnetic energy dissipation is first facilitated by the tearing instability and followed by the secondary kink instability. Particles are mostly energized inside the magnetic islands during the tearing stage due to the spatially varying electric fields produced by the outflows from reconnection. Secondary kink instability leads to additional particle acceleration. Accelerated particles are, however, observed to be thermalized quickly. The large amplitude of the vertical magnetic field resulting from the tearing modes by the secondary kink modes further help thermalizing the non-thermal particles generated from the secondary kink instability. Implications of these results for astrophysics are briefly discussed.

/pdf/particle-energization-in-3d-magnetic-reconnection-of-1x76w6h303.pdf

Particle Energization in 3D Magnetic Reconnection of Relativistic Pair Plasmas

Fully electromagnetic particle in cell simulations of nonlinear waves propagation and interaction are performed in two-dimensional plane geometry in magnetized plasma in ion cyclotron frequency range. A spectrum of fast magnetosonic wave modes with wave numbers parallel and perpendicular to the uniform equilibrium magnetic field is launched into plasma and the nonlinear dynamics of these waves is analyzed. Results show that the wave magnetic energy spectrum cascades to smaller scales. In the low frequency in the magnetohydrodynamic regime, the cascade is basically isotropic. Once entering the high frequency kinetic regime the cascade exhibits strong anisotropy, it extends to much smaller scales in direction perpendicular to the equilibrium magnetic field. The shape of the cascade is established after a few ion cyclotron periods and most of the energy in the cascade stays in the fast wave oscillations. Collisionless damping on electrons is the main dissipation channel in these results.

Particle in cell simulations of fast magnetosonic wave turbulence in the ion cyclotron frequency range

We investigate the magnetic energy transfer from the fluid to kinetic scales and dissipation processes using three-dimensional fully kinetic particle-in-cell plasma simulations. The nonlinear evolution of a sheet pinch is studied where we show that it exhibits both fluid scale global relaxation and kinetic scale collisionless reconnection at multiple resonant surfaces. The interactions among collisionless tearing modes destroy the original flux surfaces and produce stochastic fields, along with generating sheets and filaments of intensified currents. In addition, the magnetic energy is transferred from the original shear length scale both to the large scales due to the global relaxation and to the smaller, kinetic scales for dissipation. The dissipation is dominated by the thermal or pressure effect in the generalized Ohm's law, and electrons are preferentially accelerated.

Kevin J. Bowers

Papers

Desmond Performance on a Cluster of Multicore Processors

Self-organized coherent bursts of stimulated Raman scattering and speckle interaction in multi-speckled laser beams

Particle Energization in 3D Magnetic Reconnection of Relativistic Pair Plasmas

Particle in cell simulations of fast magnetosonic wave turbulence in the ion cyclotron frequency range

Spectral energy transfer and dissipation of magnetic energy from fluid to kinetic scales.