Published January 24, 2022 | Version v1
Report Open

On the potential benefits of using GPUs for atomistic MD simulations using LAMMPS

Authors/Creators

Contributors

Project leader:

Description

This report presents the performance results of classical Molecular dynamics simulations using many-body interatomic potentials for metallic systems on heterogeneous CPU-GPU architecture. Two different simulation scenarios (short(1 ps) and long (900 ps)) have been performed and com- pared for CPU only and CPU-GPU configurations to quantify GPU performance contribution. In both scenarios, GPUs made significant performance speedup by a factor of 20. Furthermore, different computational resource configurations were compared to assess the influence of commu- nication overhead, latency, proximity, and workload. The comparisons suggest an optimal choice of CPU/GPU ratio and distributed workload (load balancing) yield the best performance for the investigated simulations on heterogenous CPU-GPU architecture.

Files

LAMMPS_GPU_Report_Revised_2022.pdf

Files (447.6 kB)

Name Size Download all
md5:b870c9a12cfd81e4e842dc3e0967c199
447.6 kB Preview Download

Additional details

Related works

Cites
Journal: 10.1016/j.matdes.2021.110282 (DOI)
Journal: 10.3389/fmats.2020.602567 (DOI)
Journal: 10.1016/j.commatsci.2018.08.055 (DOI)
Thesis: 10.13154/294-6470 (DOI)
Journal article: 10.1088/1742-6596/1740/1/012001 (DOI)

Funding

Helmholtz-Zentrum Hereon
I2B project MetalMD I2B project MetalMD

References

  • Aidan P Thompson, H Metin Aktulga, Richard Berger, Dan S Bolintineanu, W Michael Brown, Paul S Crozier, Pieter J in't Veld, Axel Kohlmeyer, Stan G Moore, Trung Dac Nguyen, et al. Lammps-a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Computer Physics Communications, 271:108171, 2022.
  • Hariprasath Ganesan, Ingo Scheider, and Christian J Cyron. Understanding creep in TiAl alloys on the nanosecond scale by molecular dynamics simulations. Materials & Design, 212:110282, 2021.
  • Murray S Daw and Michael I Baskes. Embedded-atom method: Derivation and application to impurities, surfaces, and other defects in metals. Physical Review B, 29(12):6443, 1984.
  • Rajendra R Zope and Yu Mishin. Interatomic potentials for atomistic simulations of the ti-al system. Physical Review B, 68(2):024102, 2003.
  • Hariprasath Ganesan, Ingo Scheider, and Christian J Cyron. Quantifying the high-temperature separation behavior of lamellar interfaces in γ-titanium aluminide under tensile loading by molec- ular dynamics. Frontiers in Materials, 7:602567, 2021.
  • Gennady B Sushko, Alexey V Verkhovtsev, Alexander V Yakubovich, Stefan Schramm, and Andrey V Solov'yov. Molecular dynamics simulation of self-diffusion processes in titanium in bulk material, on grain junctions and on surface. The Journal of Physical Chemistry A, 118(33):6685– 6691, 2014.
  • J Fikar and R Sch ̈aublin. Molecular dynamics simulation of radiation damage in bcc tungsten. Journal of nuclear materials, 386:97–101, 2009.
  • V Yamakov, D Wolf, SR Phillpot, AK Mukherjee, and H Gleiter. Deformation-mechanism map for nanocrystalline metals by molecular-dynamics simulation. Nature materials, 3(1):43–47, 2004.
  • Cindy L Rountree, Rajiv K Kalia, Elefterios Lidorikis, Aiichiro Nakano, Laurent Van Brutzel, and Priya Vashishta. Atomistic aspects of crack propagation in brittle materials: Multimillion atom molecular dynamics simulations. Annual Review of Materials Research, 32(1):377–400, 2002.
  • Steve Plimpton. Fast parallel algorithms for short-range molecular dynamics. Journal of computational physics, 117(1):1–19, 1995.
  • Danny Perez, Blas P Uberuaga, Yunsic Shim, Jacques G Amar, and Arthur F Voter. Accelerated molecular dynamics methods: introduction and recent developments. Annual Reports in computational chemistry, 5:79–98, 2009.
  • H Ganesan, C Teijeiro, and G Sutmann. Parallelization comparison and optimization of a scale- bridging framework to model cottrell atmospheres. Computational materials science, 155:439– 449, 2018.
  • Hariprasath Ganesan. Highly parallel molecular dynamics/Monte Carlo coupling towards solutes segregation modelling. PhD Thesis, Ruhr universita ̈t Bochum, 2019.
  • Hariprasath Ganesan, Margarita Longsworth, and Godehard Sutmann. Parallel hybrid Monte Carlo/Molecular Statics for Simulation of Solute Segregation in Solids. In Journal of Physics: Conference Series, volume 1740, page 012001. IOP Publishing, 2021.
  • Steven J Plimpton and Aidan P Thompson. Computational aspects of many-body potentials. MRS bulletin, 37(5):513–521, 2012.