Significant improvements are presented for the molecular dynamics code ls1 mardyn — a linked cell-based code for simulating a large number of small, rigid molecules with application areas in chemical engineering. The changes consist of a redesign of the SIMD vectorization via wrappers, MPI improvements and a software redesign to allow memory-efficient execution with the production trunk to increase portability and extensibility. Two novel, memory-efficient OpenMP schemes for the linked cell-based force calculation are presented, which are able to retain Newton’s third law optimization. Comparisons to well-optimized Verlet list-based codes, such as LAMMPS and GROMACS, demonstrate the viability of the linked cell-based approach. The present version of ls1 mardyn is used to run simulations on entire supercomputers, maximizing the number of sampled atoms. Compared to the preceding version of ls1 mardyn on the entire set of 9216 nodes of SuperMUC, Phase 1, 27% more atoms are simulated. Weak scaling performance is increased by up to 40% and strong scaling performance by up to more than 220%. On Hazel Hen, strong scaling efficiency of up to 81% and 189 billion molecule updates per second is attained, when scaling from 8 to 7168 nodes. Moreover, a total of 20 trillion atoms is simulated at up to 88% weak scaling efficiency running at up to 1.33 PFLOPS. This represents a fivefold increase in terms of the number of atoms simulated to date.

No abstract

No abstract

No abstract

Coalescence of argon droplets with a radius of 25, 50, and 100 nm is studied with computational methods. Molecular dynamics (MD) simulations are carried out to generate reference data. Moreover, a phase-field model resting on a Helmholtz energy equation of state is devised and evaluated by computational fluid dynamics (CFD) simulations. Exactly the same scenarios in terms of geometry, fluid, and state are considered with these approaches. The MD and CFD simulation results show an excellent agreement over the entire coalescence process, including the decay of the inertia-induced oscillation of the merged droplet. Theoretical knowledge about the asymptotic behavior of coalescence process regimes is confirmed. All considered scenarios cross from the inertially limited viscous regime over to the inertial regime because of the low shear viscosity of argon. The particularly rapid dynamics during the initial stages of the coalescence process in the thermal regime is also captured by the phase-field model, where a closer look at the liquid density reveals that metastable states associated with negative pressure are attained in the emerging liquid bridge between the coalescing droplets. This demonstrates that this model is even capable of adequately handling the onset of coalescence. To speed up CFD simulations, the phase-field model is transferred to coarser grids through an interface widening approach that retains the thermodynamic properties including the surface tension.