Hprc banner tamu.png

Difference between revisions of "SW:LAMMPS"

From TAMU HPRC
Jump to: navigation, search
(added grace example)
 
(2 intermediate revisions by 2 users not shown)
Line 21: Line 21:
 
  [ netID@cluster ~]$ '''module load LAMMPS/17Nov2016-intel-2017A'''
 
  [ netID@cluster ~]$ '''module load LAMMPS/17Nov2016-intel-2017A'''
  
<font color=teal>'''NOTE:''' The normal LAMMPS program is called lmp_linux.  The GPU-enabled LAMMPS program is called lmp_gpu.'''</font>
+
<font color=teal>'''NOTE:''' On Terra, The normal LAMMPS program is called lmp_linux.  The GPU-enabled LAMMPS program is called lmp_gpu.'''</font>
 +
 
 +
<font color=teal>'''NOTE:''' On Grace, The normal LAMMPS program is called lmp. '''</font>  
  
 
{{:SW:Login_Node_Warning}}
 
{{:SW:Login_Node_Warning}}
Line 79: Line 81:
 
<pre>
 
<pre>
 
#!/bin/bash
 
#!/bin/bash
## ENVIRONMENT SETTINGS; CHANGE WITH CAUTION
 
#SBATCH --export=NONE            # Do not propagate environment
 
#SBATCH --get-user-env=L        # Replicate login environment
 
  
 
## NECESSARY JOB SPECIFICATIONS
 
## NECESSARY JOB SPECIFICATIONS
Line 102: Line 101:
 
<pre>
 
<pre>
 
#!/bin/bash
 
#!/bin/bash
## ENVIRONMENT SETTINGS; CHANGE WITH CAUTION
 
#SBATCH --export=NONE            # Do not propagate environment
 
#SBATCH --get-user-env=L        # Replicate login environment
 
  
 
## NECESSARY JOB SPECIFICATIONS
 
## NECESSARY JOB SPECIFICATIONS
Line 120: Line 116:
 
# these extra arguments enable the gpu suffix for various styles AND tell LAMMPS to use two GPUs per node
 
# these extra arguments enable the gpu suffix for various styles AND tell LAMMPS to use two GPUs per node
 
mpirun lmp_gpu -sf gpu -pk gpu 2 -in inputfile
 
mpirun lmp_gpu -sf gpu -pk gpu 2 -in inputfile
 +
</pre>
 +
 +
=== Grace Examples ===
 +
 +
On Grace, to submit a batch job, run:
 +
[NetID@grace1 ~]$ '''sbatch ''jobscript'''''
 +
 +
==== MPI Job Example ====
 +
 +
<pre>
 +
#!/bin/bash
 +
 +
## NECESSARY JOB SPECIFICATIONS
 +
#SBATCH --job-name=lammps        # Set the job name to "lammps"
 +
#SBATCH --time=01:30:00          # Set the wall clock limit to 1hr and 30min
 +
#SBATCH --ntasks=56              # Request 56 tasks
 +
#SBATCH --ntasks-per-node=28    # Request 28 task per node
 +
#SBATCH --mem=40G                # Request 40GB per node
 +
#SBATCH --output=lammps.%j      # Send stdout/err to "lammps.[jobID]"
 +
 +
ml GCC/8.3.0 OpenMPI/3.1.4
 +
ml LAMMPS/3Mar2020-Python-3.7.4-kokkos
 +
 +
mpirun lmp -in inputfile
 
</pre>
 
</pre>
  
 
[[Category:Software]]
 
[[Category:Software]]

Latest revision as of 09:24, 2 November 2021

LAMMPS

Description

LAMMPS is a classical molecular dynamics code, and an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator from Sandia National Laboratories. - Homepage: http://lammps.sandia.gov/

Documentation: http://lammps.sandia.gov/doc/Manual.html

Access

LAMMPS is open to all HPRC users.

Loading the Module

To see all versions of LAMMPS available:

[ netID@cluster ~]$ module spider LAMMPS

NOTE: HPRC typically only installs the stable releases of LAMMPS on the clusters.

To load a particular version of LAMMPS (Example: 17Nov2016-intel-2017A):

[ netID@cluster ~]$ module load LAMMPS/17Nov2016-intel-2017A

NOTE: On Terra, The normal LAMMPS program is called lmp_linux. The GPU-enabled LAMMPS program is called lmp_gpu.

NOTE: On Grace, The normal LAMMPS program is called lmp.

Usage on the Login Nodes

Please limit interactive processing to short, non-intensive usage. Use non-interactive batch jobs for resource-intensive and/or multiple-core processing. Users are requested to be responsible and courteous to other users when using software on the login nodes.

The most important processing limits here are:

  • ONE HOUR of PROCESSING TIME per login session.
  • EIGHT CORES per login session on the same node or (cumulatively) across all login nodes.

Anyone found violating the processing limits will have their processes killed without warning. Repeated violation of these limits will result in account suspension.
Note: Your login session will disconnect after one hour of inactivity.

Usage on the Compute Nodes

Non-interactive batch jobs on the compute nodes allows for resource-demanding processing. Non-interactive jobs have higher limits on the number of cores, amount of memory, and runtime length.

For instructions on how to create and submit a batch job, please see the appropriate wiki page for each respective cluster:

Ada Examples

On Ada, to submit a batch job, run:

[NetID@ada1 ~]$ bsub < jobscript

MPI Job Example

#!/bin/bash

#BSUB -J lammps                   # Set the job name to "lammps"
#BSUB -W 1:00                     # Set the wall clock limit to 1hr
#BSUB -n 40                       # Request 40 tasks 
#BSUB -R "span[ptile=20]"         # Request 20 tasks per node
#BSUB -M 2500                     # Set effective memory limit to 2500 MB per task
#BSUB -R "rusage[mem=2500]"       # Reserve 2500 MB of memory per task
#BSUB -o lammps.%J                # Send stdout/err to "lammps.[jobID]"

ml LAMMPS/17Nov2016-intel-2017A

mpirun -np 40 lmp_linux -in in.input

GPU Job Example

#!/bin/bash

#BSUB -J lammps_gpu               # Set the job name to "lammps_gpu"
#BSUB -W 1:00                     # Set the wall clock limit to 1hr
#BSUB -n 40                       # Request 40 tasks 
#BSUB -R "span[ptile=20]"         # Request 20 tasks per node
#BSUB -M 2500                     # Set effective memory limit to 2500 MB per task
#BSUB -R "rusage[mem=2500]"       # Reserve 2500 MB of memory per task
#BSUB -o lammps_gpu.%J            # Send stdout/err to "lammps_gpu.[jobID]"
#BSUB -R "select[gpu]"            # Use only gpu nodes for this job

ml LAMMPS/17Nov2016-intel-2017A

# these extra arguments enable the gpu suffix for various styles AND tell LAMMPS to use two GPUs per node
mpirun -np 40 lmp_gpu -sf gpu -pk gpu 2 -in in.input

Terra Examples

On Terra, to submit a batch job, run:

[NetID@terra1 ~]$ sbatch jobscript

MPI Job Example

#!/bin/bash

## NECESSARY JOB SPECIFICATIONS
#SBATCH --job-name=lammps        # Set the job name to "lammps"
#SBATCH --time=01:30:00          # Set the wall clock limit to 1hr and 30min
#SBATCH --ntasks=56              # Request 56 tasks
#SBATCH --ntasks-per-node=28     # Request 28 task per node
#SBATCH --mem=40G                # Request 40GB per node
#SBATCH --output=lammps.%j       # Send stdout/err to "lammps.[jobID]"

ml LAMMPS/17Nov2016-intel-2017A

mpirun lmp_linux -in inputfile

GPU Job Example

NOTE: GPU jobs should be submitted to the gpu partition and MUST specify how many GPUs are required per node.

#!/bin/bash

## NECESSARY JOB SPECIFICATIONS
#SBATCH --job-name=lammps_gpu    # Set the job name to "lammps_gpu"
#SBATCH --time=01:30:00          # Set the wall clock limit to 1hr and 30min
#SBATCH --ntasks=56              # Request 56 tasks
#SBATCH --ntasks-per-node=28     # Request 28 task per node
#SBATCH --mem=40G                # Request 40GB per node
#SBATCH --output=lammps_gpu.%j   # Send stdout/err to "lammps_gpu.[jobID]"
#SBATCH --gres=gpu:2             # Request 2 GPUs per node
#SBATCH --partition=gpu          # Submit job to the gpu queue

ml LAMMPS/17Nov2016-intel-2017A

# these extra arguments enable the gpu suffix for various styles AND tell LAMMPS to use two GPUs per node
mpirun lmp_gpu -sf gpu -pk gpu 2 -in inputfile

Grace Examples

On Grace, to submit a batch job, run:

[NetID@grace1 ~]$ sbatch jobscript

MPI Job Example

#!/bin/bash

## NECESSARY JOB SPECIFICATIONS
#SBATCH --job-name=lammps        # Set the job name to "lammps"
#SBATCH --time=01:30:00          # Set the wall clock limit to 1hr and 30min
#SBATCH --ntasks=56              # Request 56 tasks
#SBATCH --ntasks-per-node=28     # Request 28 task per node
#SBATCH --mem=40G                # Request 40GB per node
#SBATCH --output=lammps.%j       # Send stdout/err to "lammps.[jobID]"

ml GCC/8.3.0 OpenMPI/3.1.4
ml LAMMPS/3Mar2020-Python-3.7.4-kokkos

mpirun lmp -in inputfile