591 TFLOPS Multi-trillion Particles Simulation on SuperMUC

Eckhardt, Wolfgang; Heinecke, Alexander; Bader, Reinhold; Brehm, Matthias; Hammer, N. J.; Huber, Herbert; Kleinhenz, Hans-Georg; Vrabec, Jadran; Hasse, Hans; Horsch, Martin; Bernreuther, Martin; Glass, Colin W.; Niethammer, Christoph; Bode, Arndt; Bungartz, Hans‐Joachim

doi:10.1007/978-3-642-38750-0_1

Cited by 51 publications

(51 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A version of ls1 mardyn was optimized for simulating single-site LJ particles on the SuperMUC system, 21 one of the largest x86 systems worldwide with 147 500 cores and a theoretical peak performance of more than 3 PFLOPS. It is based on a high-performance FDR-10 InfiniBand inter- To evaluate the performance with respect to strong scaling behavior, a scenario with N = 9.5 × 10 8 particles was studied, which fits into the memory of two nodes, as 18 GB per node are needed.…”

Section: Trillion Particle Simulationmentioning

confidence: 99%

“…As described by Eckhardt et al, 21 a larger weak scaling benchmark on the whole SuperMUC was performed with that code version. Simulating 4.125 × 10 12 molecules, to our knowledge the largest MD simulation to date, with a cut-off radius of r c = 3.5 σ , one time step took roughly 40 s.…”

Section: Trillion Particle Simulationmentioning

confidence: 99%

“…of the order of 10 23 , the largest ensembles that can be handled today are of the order of 10 12 molecules. 21 This limitation is usually addressed by focusing on representative subvolumes, containing a limited number of molecules, to which an appropriate set of boundary conditions is applied. Depending on the type of information that is determined, e.g.…”

Section: Introductionmentioning

confidence: 99%

“…23,51,52 On the SuperMUC, over four trillion molecules were simulated. 21 The vapor-liquid surface tension and its dependence on size and curvature was characterized. 49,[53][54][55][56] The ls1 mardyn program was furthermore employed to investigate fluid flow through nanoporous membrane materials 57 and adsorption phenomena such as the fluid-solid contact angle in dependence on the fluid-wall interaction.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

ls1 mardyn: The Massively Parallel Molecular Dynamics Code for Large Systems

Niethammer

Becker

Bernreuther

et al. 2014

J. Chem. Theory Comput.

Self Cite

121

View full text Add to dashboard Cite

The molecular dynamics simulation code ls1 mardyn is presented. It is a highly scalable code, optimized for massively parallel execution on supercomputing architectures, and currently holds the world record for the largest molecular simulation with over four trillion particles. It enables the application of pair potentials to length and time scales which were previously out of scope for molecular dynamics simulation. With an efficient dynamic load balancing scheme, it delivers high scalability even for challenging heterogeneous configurations. * To whom correspondence should be addressed † High Performance Computing Center Stuttgart (HLRS), Germany ‡ Laboratory of Engineering Thermodynamics (LTD), Univ. of Kaiserslautern, Germany ¶ Scientific Computing in Computer Science (SCCS), TU München, Germany § Thermodynamics and Energy Technology (ThEt), Univ. of Paderborn, Germany 1 Presently, multi-center rigid potential models based on Lennard-Jones sites, point charges and higher-order polarities are supported. Due to its modular design, ls1 mardyn can be extended to new physical models, methods, and algorithms, allowing future users to tailor it to suit their respective needs. Possible applications include scenarios with complex geometries, e.g.for fluids at interfaces, as well as non-equilibrium molecular dynamics simulation of heat and mass transfer.

show abstract

Section: Trillion Particle Simulationmentioning

confidence: 99%

Section: Trillion Particle Simulationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

ls1 mardyn: The Massively Parallel Molecular Dynamics Code for Large Systems

Niethammer

Becker

Bernreuther

et al. 2014

J. Chem. Theory Comput.

Self Cite

121

View full text Add to dashboard Cite

show abstract

“…For example, a single barely-visible droplet of liquid contains on the order of a trillion molecules. Sim-ulations on such a scale are costly, but have become feasible [7,22]. This enables comparison with human-observable experiments.…”

Section: Introductionmentioning

confidence: 99%