hpc documentation

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== LAMMPS on Caviness ======

A Slurm job submission script template for batch LAMMPS jobs is available at ''/opt/shared/templates/slurm/applications/lammps.qs''.  Copy and edit the template based on your job's resource needs; the comments in the template describe each Slurm option and environment variable present.

<note tip>It is a good idea to periodically check in ''/opt/shared/templates/slurm/applications'' for changes in existing templates, or the addition of new templates, designed to provide the best performance for a particular version of LAMMPS on that cluster.</note>


At the very least, the LAMMPS input file variable must be altered accordingly:

<code bash>
#
# [EDIT] Indicate your LAMMPS input file (mandatory)
#
LAMMPS_INPUT="in.indent"
</code>

Also keep in mind, the default version of LAMMPS is loaded via VALET in the job submission script template and the default version may change over time based on new releases and bug fixes that indicate an older version should no longer be used.  Below is the VALET command specified in the job submission script template

<code bash>
#
# [EDIT] Load some version of LAMMPS into the job environment
#
vpkg_require lammps/default
</code>

Based on the versions of LAMMPS available from ''vpkg_versions lammps'', we might see something like this

<code bash>
$ vpkg_versions lammps

Available versions in package (* = default version):

[/opt/shared/valet/2.1/etc/lammps.vpkg_yaml]
lammps                 Molecular Dynamics Simulator
  12Dec2018:intel,mkl  release 12Dec2018 (Intel compilers, MPI, MKL FFT, USER-INTEL, USER-OMP)
  13Mar2018            release 13Mar2018 (system compilers, MPI, FFTW3)
* 13Mar2018:intel      release 13Mar2018 (Intel compilers, MPI, FFTW3)
  13Mar2018:intel,mkl  release 13Mar2018 (Intel compilers, MPI, MKL FFT)
  28Jun2014            release 28Jun2014 (system compilers, MPI, FFTW3)
  28Jun2014:intel      release 28Jun2014 (Intel compilers, MPI, FFTW3)
  28Jun2014:intel,mkl  release 28Jun2014 (Intel compilers, MPI, MKL FFT)
</code>

 if we want to specify the ''28Jun2014:intel,mkl'' instead of the default version (''13Mar2018:intel'') indicated by the ''*'', then we would change the job submission script to the following

<code bash>
#
# [EDIT] Load some version of LAMMPS into the job environment
#
vpkg_require lammps/28Jun2014:intel,mkl
</code>

The job submission script template sources an external script to complete the setup of the job environment. You should never delete parts of the template or comment out any option unless you see ''[EDIT]'' instructions.

<note warning>**Remember:** The clusters may have a few versions of a lammps package available but users are cautioned: they contain only the default functionality with which the program ships. Any additional functionalities realized by enabling packages require end users (or workgroups) to build their own copies of LAMMPS. Please reference [[technical:recipes:lammps|Installing and managing LAMMPS on Caviness/DARWIN]] as a guide through this process.</note>

<note important>There have been a few reports from clients that after a LAMMPS job has been submitted (especially if you know the job has worked correctly before), the scheduler indicates the job is running however no data is written in the output files. In such cases, clients report cancelling the job and resubmitting without any change and the job ran successfully.</note>