Стегайлов Владимир Владимирович

Performance and portability of state-of-art molecular dynamics software on modern GPUs

2020 · CHAPTER · en

Classical molecular dynamics (MD) calculations represent a significant part of utilization time of high performance computing systems. As usual, efficiency of such calculations is based on an interplay of software and hardware that are nowadays moving to hybrid GPU-based technologies. Several well-developed MD packages focused on GPUs differ both in their data management capabilities and in performance. In this paper, we present our results for the porting of the CUDA backend of LAMMPS to ROCm HIP that shows considerable benefits for AMD GPUs comparatively to the existing OpenCL backend. We consider the efficiency of solving the same physical models using different software and hardware combinations. We analyze the performance of LAMMPS, HOOMD, GROMACS and OpenMM MD packages with different GPU back-ends on modern Nvidia Volta and AMD Vega20 GPUs.

Nonadiabatic effects and excitonlike states during the insulator-to-metal transition in warm dense hydrogen

2020 · ARTICLE · en

The transition of molecular hydrogen to atomic ionized state with the increase of temperature and pressure poses still unresolved problems for experimental methods and theory. Here we analyze the dynamics of this transition and show its nonequilibrium nonadiabatic character overlooked in both interpreting experimental data and in theoretical models. The nonadiabatic mechanism explains the strong isotopic effect [M. Zaghoo, R. J. Husband, and I. F. Silvera, Phys. Rev. B 98, 104102 (2018).] and the large latent heat [M. Houtput, J. Tempere, and I. F. Silvera, Phys. Rev. B 100, 134106 (2019).] reported recently. We demonstrate the possibility of the formation of intermediate excitonlike molecular states at heating of molecular hydrogen that can explain the puzzling experimental data on reflectivity and conductivity during the insulator-to-metal transition.

Nanobubbles diffusion in bcc uranium: Theory and atomistic modelling

2020 · ARTICLE · en

The presence of stable facets of nanobubbles in crystal lattice can significantly affect their diffusion coefficient, but the existing theory of this phenomenon is too general and can not take into account atomistic structure of nanobubbles in a given material. Such a theory for the mechanisms of bubble motion in crystals can be extended and developed using methods of atomistic modelling. In this work, we consider the movement of bubbles in the bcc lattice of gamma-U. The Beere’s theory of faceted bubble motion is revised and a method of non-equilibrium accelerated molecular dynamics in a pressure gradient is proposed. The results of the accelerated method calculations for gamma-U are verified using generic molecular-dynamics calculations of free nanobubble diffusion. The new method significantly accelerates calculations of the diffusion coefficient for nanometer-sized bubbles and opens the way for more accurate material-specific calculations of gas-filled nanobubbles diffusivity in nuclear fuels.

High temperature pure carbon nanoparticle formation: Validation of AIREBO and ReaxFF reactive molecular dynamics

2020 · ARTICLE · en

Molecular dynamics with reactive interatomic potentials is the only computationally feasible approach for modeling at the atomistic level the formation of carbon nanoparticles from gas state. Such models require thousands of atoms and millions of time steps that is beyond the current capabilities of first principles electronic structure calculations. A continuously growing variety of available reactive interatomic potentials for carbon requires their careful validation for a particular molecular system and pressure-temperature conditions. In this work we consider a generic example process of carbon nanoparticle formation at cooling of the gas phase and compare different AIREBO and ReaxFF reactive models. Three main processes of clusterization, change of hybridization and graphitization are analysed and used for comparison of potentials. Ab initio density functional theory and parameterized density functional tight-binding calculations together with experimental data available are used for validation of the reactive models considered.We highlight the detected problems of some well-known reactive potentials and conclude with three models that can be selected as the best options for molecular dynamics simulations of pure carbon nanoparticle formation.

Matrix-Matrix Multiplication Using Multiple GPUs Connected by Nvlink

2020 · CHAPTER · en

Angara interconnect makes GPU-based Desmos supercomputer an efficient tool for molecular dynamics calculations

2019 · ARTICLE · en

In this paper, we describe the Desmos supercomputer that consists of 32 hybrid nodes connected by a low-latency highbandwidth Angara interconnect with torus topology. This supercomputer is aimed at cost-effective classical molecular dynamics calculations. Desmos serves as a test bed for the Angara interconnect that supports 3D and 4D torus network topologies, and verifies its ability to unite massively-parallel programming systems speeding-up effectively MPI-based applications. We describe the Angara interconnect presenting typical MPI benchmarks. Desmos benchmarks results for GROMACS, LAMMPS, VASP and CP2K are compared with the data for other HPC systems. Also, we consider the job scheduling statistics for several months of Desmos deployment.

Reactive molecular-dynamics study of onion-like carbon nanoparticle formation

2019 · ARTICLE · en

Formation of carbon nanoparticles is an important type of complex non-equilibrium processes that require precise atomistic theoretical understanding. In this work, we consider the process of ultrafast cooling of pure carbon gas that results in nucleation of an onion-like fullerene. The model is based on molecular dynamics simulation with the interaction between carbon atoms described via a reactive ReaxFF model. We study the consecutive stages of fullerene-like nanoparticle formation and identify the corresponding temperature ranges. Analysis of hybridization and graphitization reveals the underlying microscopic mechanisms connected with rearrangements of dihedral angles and density changes.

Deploying Elbrus VLIW CPU Ecosystem for Materials Science Calculations: Performance and Problems.

2019 · CHAPTER · en

Modern Elbrus-4S and Elbrus-8S processors show floating point performance comparable to the popular Intel processors in the field of high-performance computing. Tasks oriented to take advantage of the VLIW architecture show even greater efficiency on Elbrus processors. In this paper the efficiency of the most popular materials science codes in the field of classical molecular dynamics and quantum-mechanical calculations is considered. A comparative analysis of the performance of these codes on Elbrus processor and other modern processors is carried out.

Formation free energies of point defects and thermal expansion of bcc U and Mo

2019 · ARTICLE · en

Performance and Scalability of Materials Science and Machine Learning Codes on the State-of-Art Hybrid Supercomputer Architecture

2019 · CHAPTER · en

8 of top 10 supercomputers of Top500 list published in November 2018 consist of computing nodes with hybrid architectures that require special programming techniques. 5 systems among these are based on Nvidia GPUs. In this paper, we consider the benchmark results of the brand new hybrid supercomputer installed in March 2019 in NRU HSE. This system gives us the possibility to estimate the performance of several widely used material science and machine learning codes that we discuss in this work within the framework of the results available for older HPC systems.