Polymer in water tutorial files not working on the HPC

LAMMPS LAMMPS Beginners
1

Hi All,

I am working through the polymer in water tutorial from the github and have ran into a problem. Working locally there is no problems with any of the input files, but switching to the HPC (1 node and 1 core) the same file throws an Out of range atoms - cannot compute PPPM error code during the 20,000 steps of NVT after minimization of the pure water box. What about switching to the HPC version would cause the error to occur and/or what am I missing in my understanding of PPPM/LAMMPS?

Any input would be super helpful,
Kyle

2

GitHub is large and has many repositories and quite a few are tutorials. Which specific repository are you referring to?

Hard to say without knowing which versions of LAMMPS you are using locally and remotely and what specific platforms you are running on.

You don’t even say exactly how you are running on the HPC system?

Unless you provide more specific information that allows to reproduce what you are seeing, it will be extremely difficult to give any specific advice.

3

My apologies for being vague. I am working with the LAMMPS tutorial by Simon Gravelle called Polymer in Water and I am stuck on the Preparing the water reservoir section.

Locally I am using the most current release for MAC of LAMMPS with only the default packages installed.

Remotely I am using LAMMPS version from 24 Dec 2020 complied to use Open MPI for parallelization for use in Linux (bash).

Flags for Open MPI

– mpi job startup = /opt/shared/openmpi/3.1.3-intel/bin/mpirun
– nhosts = 1
– nproc = 1
– nproc-per-node = 1
– cpus-per-proc = 1

Open MPI environment flags:
– OMPI_MCA_btl_base_exclude=tcp
– OMPI_MCA_rmaps_base_display_map=true
– OMPI_MCA_orte_hetero_nodes=true
– OMPI_MCA_hwloc_base_binding_policy=core
– OMPI_MCA_rmaps_base_mapping_policy=core

The command line to excute the code is /opt/shared/openmpi/3.1.3-intel/bin/mpirun lmp -in input.lammps

hope this gives better context for the error message

4

Hello,

These tutorials are written for the 2Aug2023 version, and using such an old version may trigger additional error.

On the side note, I just updated the polymer-in-water tutorial. I corrected the values of the bond coefficients, and changed the LJ cut-off to a smaller value of 1 nm (which is closer to the value used in the original water model article 10.1063/1.2136877). Since then, I ran the simulations with several different SEEDs and did not see the Out of range atoms - cannot compute PPPM error again.

Best,
Simon

2 Likes