/
Getting Started with Secret App LAMMPS for ISC25 SCC (In-Person)

Getting Started with Secret App LAMMPS for ISC25 SCC (In-Person)

Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) is a classical molecular dynamics code. LAMMPS has potentials for solid-state materials (metals, semiconductors) and soft matter (biomolecules, polymers) and coarse-grained or mesoscopic systems. It can be used to model atoms or, more generically, as a parallel particle simulator at the atomic, meso, or continuum scale. LAMMPS runs on single processors or in parallel using message-passing techniques and a spatial-decomposition of the simulation domain. LAMMPS is distributed as an open source code under the terms of the GPL. More information on LAMMPS can be found at the LAMMPS web site: http://lammps.sandia.gov.

 

 

Introduction to LAMMPS

LAMMPS is a classical molecular dynamics code with a focus on materials modeling. It's an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator.

LAMMPS has potentials for solid-state materials (metals, semiconductors) and soft matter (biomolecules, polymers) and coarse-grained or mesoscopic systems. It can be used to model atoms or, more generically, as a parallel particle simulator at the atomic, meso, or continuum scale.

Build LAMMPS for x86_64 CPUs

This is an example that uses HPC-X MPI.

  1. Load the relevant modules

module load intel/2023.2 module load compiler mkl module load hpcx-2.22.1

2. Make

# Download latest stable version from https://lammps.sandia.gov/download.html. tar xf lammps-stable.tar.gz cd lammps-29Aug2024/src make clean-all make no-all make no-lib make yes-intel yes-kspace yes-molecule yes-rigid yes-manybody make -j 32 intel_cpu_openmpi

  1. Run example

mpirun -np 128 lammps/lmp_intel_cpu_openmpi -in <input> -pk intel 0 omp 1 -sf intel

Build LAMMPS for ARM CPUs

This is an example that uses nvhpc on Nvidia Grace CPU.

Download and install nvhpc for ARM from https://developer.nvidia.com/hpc-sdk-downloads.

  1. Load the relevant modules

module use <path to nvhpc-25.3>/modulefiles module load nvhpc-hpcx/25.3

  1. Make

# Download latest stable version from https://lammps.sandia.gov/download.html. tar xf lammps-stable.tar.gz cd lammps-29Aug2024/src cd MAKE/OPTIONS cp Makefile.g++_openmpi Makefile.arm # Makefile.arm as needed cd ../.. make clean-all make no-all make no-lib make yes-kspace yes-molecule yes-rigid yes-manybody make -j 32 arm

  1. Run example

mpirun -np 128 lammps/lmp_arm -in <input>

Output example

... Precision: mixed ---------------------------------------------------------- Neighbor list info ... update: every = 1 steps, delay = 5 steps, check = yes max neighbors/atom: 2000, page size: 100000 master list distance cutoff = 4.2 ghost atom cutoff = 4.2 binsize = 2.1, bins = 828 828 1656 1 neighbor lists, perpetual/occasional/extra = 1 0 0 (1) pair tersoff/intel, perpetual attributes: full, newton on, intel pair build: full/bin/intel stencil: full/bin/3d/intel bin: intel Setting up Verlet run ... Unit style : metal Current step : 0 Time step : 0.001 Per MPI rank memory allocation (min/avg/max) = 2.016e+04 | 2.025e+04 | 2.027e+04 Mbytes Step Temp E_pair E_mol TotEng Press 0 1000 -2.427669e+09 0 -2.3598995e+09 7019.8698 100 428.29904 -2.3888869e+09 0 -2.3598612e+09 -14372.049 200 543.33687 -2.3966807e+09 0 -2.359859e+09 -1846.3798 300 493.20805 -2.3932791e+09 0 -2.3598547e+09 -3963.2901 400 496.80952 -2.3935222e+09 0 -2.3598537e+09 -2873.4339 500 483.83489 -2.3926424e+09 0 -2.3598531e+09 -2995.8522 600 469.52456 -2.3916726e+09 0 -2.3598531e+09 -4801.5372 Loop time of 1259.88 on 224 procs for 600 steps with 524288000 atoms Performance: 0.041 ns/day, 583.276 hours/ns, 0.476 timesteps/s, 249.685 Matom-step/s 99.5% CPU use with 224 MPI tasks x 1 OpenMP threads MPI task timing breakdown: Section | min time | avg time | max time |%varavg| %total --------------------------------------------------------------- Pair | 1046 | 1050.7 | 1056.9 | 6.9 | 83.40 Neigh | 24.737 | 24.911 | 25.547 | 2.7 | 1.98 Comm | 42.334 | 46.967 | 58.881 | 45.4 | 3.73 Output | 0.14938 | 0.1639 | 0.21684 | 2.3 | 0.01 Modify | 94.349 | 102.09 | 104.56 | 19.4 | 8.10 Other | | 35.04 | | | 2.78 Nlocal: 2.34057e+06 ave 2.34273e+06 max 2.33676e+06 min Histogram: 11 49 4 0 0 0 26 38 12 84 Nghost: 179410 ave 183845 max 176690 min Histogram: 96 0 45 19 0 0 0 0 0 64 Neighs: 0 ave 0 max 0 min Histogram: 224 0 0 0 0 0 0 0 0 0 FullNghs: 3.85404e+07 ave 3.85804e+07 max 3.84722e+07 min Histogram: 3 48 12 1 0 0 43 21 52 44 Total # of neighbors = 8.6330451e+09 Ave neighs/atom = 16.466227 Neighbor list builds = 26 Dangerous builds = 0

For GPUs

https://docs.lammps.org/Speed_gpu.html

Tasks and Submissions

  1. Run LAMMPS with .

  2. Submit the standard output