Tutorial 8: LAMMPS Input Files — Building Your First Simulation

Objective: Learn to write LAMMPS input scripts by building a simple Argon gas simulation from scratch.

1. Getting Started: Creating Your Workspace¶

Before we write any LAMMPS code, let’s set up a proper workspace on Anvil.

Step 1: Log in to Anvil¶

Open a terminal on Anvil:

• Via SSH: Use your terminal application

• Via OnDemand: Go to ondemand.anvil.rcac.purdue.edu → Clusters → Anvil Shell Access

Step 2: Navigate to Your Scratch Space¶

cd $SCRATCH

Step 3: Create a Tutorial Directory¶

mkdir lammps_tutorial
cd lammps_tutorial
pwd

The pwd command shows your current location. You should see:

/anvil/scratch/x-username/lammps_tutorial

Step 4: Create Your First LAMMPS Input File¶

We’ll use the nano text editor (simple and beginner-friendly):

nano argon.in

This opens a blank file. You’ll see a text editor with a menu at the bottom.

Keep nano open — we’ll fill it with LAMMPS commands in the next section!

2. What is LAMMPS?¶

LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is one of the most widely-used molecular dynamics codes in the world. Unlike our Python toy models, LAMMPS is production-grade software designed to:

• Simulate millions of atoms efficiently

• Run in parallel across thousands of CPU cores

• Handle complex force fields (proteins, polymers, metals, etc.)

• Integrate with analysis tools and visualization software

LAMMPS is command-driven. You write a text file (the “input script”) that tells LAMMPS what to do, step by step. This tutorial will teach you the structure of that file.

2. What is LAMMPS?¶

LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) is production-grade MD software used worldwide. Unlike our Python toy model, LAMMPS can:

• Simulate millions of atoms efficiently

• Run in parallel across thousands of cores

• Handle complex materials (proteins, polymers, metals)

How it works: You write a text file (input script) telling LAMMPS what to do, step by step.

3. Building Your Input File: Section by Section¶

Now let’s fill in the argon.in file you have open in nano. Copy each section below.

The Six-Section Structure¶

Every LAMMPS input follows this template:

# SECTION 1: INITIALIZATION - Define the "universe"
# SECTION 2: SYSTEM DEFINITION - Create atoms
# SECTION 3: FORCE FIELD - Define interactions
# SECTION 4: SIMULATION SETTINGS - Timestep, thermostat
# SECTION 5: OUTPUT - What to save
# SECTION 6: EXECUTION - Run!

4. Section 1: Initialization¶

Copy this into nano:

# ========================================
# SECTION 1: INITIALIZATION
# ========================================
units           real
atom_style      atomic
boundary        p p p

What it does:

• units real — Use real-world units: Angstroms (Å), femtoseconds (fs), kcal/mol, Kelvin (K)

• atom_style atomic — Simple atoms (no bonds or molecules)

• boundary p p p — Periodic boundaries (atoms wrap around edges)

5. Section 2: System Definition¶

Add this to your file:

# ========================================
# SECTION 2: SYSTEM DEFINITION
# ========================================
# FCC lattice with 3.8 Angstrom conventional cell
lattice         fcc 3.8
# Simulation box: 8 lattice units per edge
region          simbox block 0 8 0 8 0 8
create_box      1 simbox
create_atoms    1 region simbox
mass            1 39.948
velocity        all create 500.0 87287 dist gaussian

What it does:

• Creates a simulation box with ~2048 Argon atoms

• lattice fcc 3.8 — FCC lattice with conventional cell size 3.8 Å

• region simbox block 0 8 0 8 0 8 — Define a region extending 8 lattice units

• create_box 1 simbox — Create simulation box with 1 atom type

• create_atoms 1 region simbox — Fill region with atoms on FCC lattice

• mass 1 39.948 — Argon atomic mass (amu)

• Gives atoms random velocities for T = 500 K

6. Section 3: Force Field¶

Add this:

# ========================================
# SECTION 3: FORCE FIELD
# ========================================
pair_style      lj/cut 12.0
pair_coeff      1 1 0.238 3.405
neighbor        2.0 bin
neigh_modify    every 1 delay 0 check yes

What it does:

• lj/cut 12.0 — Lennard-Jones potential with cutoff at 12.0 Å

• pair_coeff 1 1 0.238 3.405 — For Argon:

  • ε = 0.238 kcal/mol (well depth)

  • σ = 3.405 Å (atom diameter)

• neighbor 2.0 — Build neighbor list with 2.0 Å buffer

7. Section 4: Simulation Settings¶

Add this:

# ========================================
# SECTION 4: SIMULATION SETTINGS
# ========================================
timestep        2.0
fix             1 all nvt temp 500.0 500.0 100.0

What it does:

• timestep 2.0 — Time advances in steps of 2.0 fs (femtoseconds)

• fix 1 all nvt temp 500.0 500.0 100.0 — Nosé-Hoover thermostat at 500 K

  • Start temp: 500 K

  • End temp: 500 K

  • Damping: 100 fs

8. Section 5: Output¶

Add this:

# ========================================
# SECTION 5: OUTPUT
# ========================================
thermo_style    custom step temp press pe ke etotal density
thermo          500
dump            1 all custom 1000 argon.lammpstrj id type x y z vx vy vz
dump_modify     1 sort id
log             thermo.out

What it does:

• thermo 500 — Print thermodynamic data every 500 steps (= 1 ps)

• dump — Save atom positions/velocities every 1000 steps (= 2 ps)

• log thermo.out — Write output to file (for later analysis)

9. Section 6: Execution¶

Final section:

# ========================================
# SECTION 6: EXECUTION
# ========================================
run             50000
write_data      final_state.data

What it does:

• run 50000 — Simulate for 50,000 timesteps = 100 ps (picoseconds)

• write_data — Save final configuration (for restart/continuation)

10. Save and Exit¶

Now save your file:

1. Press Ctrl+O (that’s the letter O, not zero)

2. Press Enter to confirm filename

3. Press Ctrl+X to exit nano

Verify your file was created:

ls -lh argon.in

You should see:

-rw-r--r-- 1 username group 1.2K date argon.in

11. Creating an Organized Folder Structure¶

Professional MD simulations require good organization. Let’s create a proper folder structure for running multiple simulations.

Why Folder Structure Matters¶

When you run multiple simulations (different temperatures, ensembles, parameters), you need:

• ✅ Easy to find specific results

• ✅ No accidentally overwriting files

• ✅ Clear documentation of what each simulation does

The Standard Structure¶

We’ll create this hierarchy:

lammps_tutorial/
├── NVT_simulations/
│   ├── T_300K/
│   ├── T_500K/
│   └── T_700K/
└── NPT_simulations/
    ├── P_1atm/
    └── P_10atm/

Step 1: Create the Main Directories¶

cd $SCRATCH/lammps_tutorial

# Create top-level folders
mkdir -p NVT_simulations
mkdir -p NPT_simulations

# Verify
ls -lh

You should see:

drwxr-xr-x  NVT_simulations/
drwxr-xr-x  NPT_simulations/
-rw-r--r--  argon.in

Step 2: Create Temperature Subfolders (NVT)¶

cd NVT_simulations

# Create three temperature folders
mkdir T_300K T_500K T_700K

# Check
ls

Output:

T_300K  T_500K  T_700K

Step 3: Copy Input Files for Each Temperature¶

We’ll create three versions of the input file, one per temperature.

For T_300K:

cd T_300K
nano argon_300K.in

Paste this complete input file:

# ========================================
# Argon NVT Simulation at 300 K
# ========================================

# ========================================
# SECTION 1: INITIALIZATION
# ========================================
units           real
atom_style      atomic
boundary        p p p

# ========================================
# SECTION 2: SYSTEM DEFINITION
# ========================================
lattice         fcc 3.8
region          simbox block 0 8 0 8 0 8
create_box      1 simbox
create_atoms    1 region simbox
mass            1 39.948
velocity        all create 300.0 87287 dist gaussian

# ========================================
# SECTION 3: FORCE FIELD
# ========================================
pair_style      lj/cut 12.0
pair_coeff      1 1 0.238 3.405
neighbor        2.0 bin
neigh_modify    every 1 delay 0 check yes

# ========================================
# SECTION 4: SIMULATION SETTINGS
# ========================================
timestep        2.0
fix             1 all nvt temp 300.0 300.0 100.0

# ========================================
# SECTION 5: OUTPUT
# ========================================
thermo_style    custom step temp press pe ke etotal density
thermo          500
dump            1 all custom 1000 argon_300K.lammpstrj id type x y z vx vy vz
dump_modify     1 sort id
log             thermo_300K.out

# ========================================
# SECTION 6: EXECUTION
# ========================================
run             50000
write_data      final_300K.data

Save: Ctrl+O, Enter, Ctrl+X

For T_500K:

cd ../T_500K
nano argon_500K.in

Paste the same file BUT change these lines:

# Header comment: "Argon NVT Simulation at 500 K"
velocity        all create 500.0 87287 dist gaussian
fix             1 all nvt temp 500.0 500.0 100.0
dump            1 all custom 1000 argon_500K.lammpstrj id type x y z vx vy vz
log             thermo_500K.out
write_data      final_500K.data

For T_700K:

cd ../T_700K
nano argon_700K.in

Change to 700 K:

# Header comment: "Argon NVT Simulation at 700 K"
velocity        all create 700.0 87287 dist gaussian
fix             1 all nvt temp 700.0 700.0 100.0
dump            1 all custom 1000 argon_700K.lammpstrj id type x y z vx vy vz
log             thermo_700K.out
write_data      final_700K.data

Step 4: Verify NVT Folder Structure¶

cd $SCRATCH/lammps_tutorial/NVT_simulations
tree

Or without tree:

ls -R

You should see:

T_300K/:
argon_300K.in

T_500K/:
argon_500K.in

T_700K/:
argon_700K.in

12. Creating NPT Simulation Inputs¶

NPT simulations control both temperature AND pressure. Let’s create two pressure conditions.

Step 1: Create Pressure Subfolders¶

cd $SCRATCH/lammps_tutorial/NPT_simulations

mkdir P_1atm P_10atm

ls

Output:

P_1atm  P_10atm

Step 2: NPT Input for 1 atm¶

cd P_1atm
nano argon_npt_1atm.in

Paste this:

# ========================================
# Argon NPT Simulation at 1 atm
# T = 500 K, P = 1 atm
# ========================================

# ========================================
# INITIALIZATION
# ========================================
units           real
atom_style      atomic
boundary        p p p

# ========================================
# SYSTEM DEFINITION
# ========================================
lattice         fcc 3.8
region          simbox block 0 8 0 8 0 8
create_box      1 simbox
create_atoms    1 region simbox
mass            1 39.948
velocity        all create 500.0 87287 dist gaussian

# ========================================
# FORCE FIELD
# ========================================
pair_style      lj/cut 12.0
pair_coeff      1 1 0.238 3.405
neighbor        2.0 bin
neigh_modify    every 1 delay 0 check yes

# ========================================
# SIMULATION SETTINGS (NPT!)
# ========================================
timestep        2.0
fix             1 all npt temp 500.0 500.0 100.0 iso 1.0 1.0 1000.0

# ========================================
# OUTPUT
# ========================================
thermo_style    custom step temp press pe ke etotal vol density
thermo          500
dump            1 all custom 2000 argon_npt_1atm.lammpstrj id type x y z
dump_modify     1 sort id
log             thermo_npt_1atm.out

# ========================================
# EXECUTION
# ========================================
run             100000
write_data      final_npt_1atm.data

Step 3: NPT Input for 10 atm¶

cd ../P_10atm
nano argon_npt_10atm.in

Copy the 1 atm file but change:

# Header: "Argon NPT Simulation at 10 atm"
# T = 500 K, P = 10 atm
fix             1 all npt temp 500.0 500.0 100.0 iso 10.0 10.0 1000.0
dump            1 all custom 2000 argon_npt_10atm.lammpstrj id type x y z
log             thermo_npt_10atm.out
write_data      final_npt_10atm.data

Step 4: Verify Complete Structure¶

cd $SCRATCH/lammps_tutorial
tree

Or:

find . -name "*.in" -type f

You should see all 5 input files:

./NVT_simulations/T_300K/argon_300K.in
./NVT_simulations/T_500K/argon_500K.in
./NVT_simulations/T_700K/argon_700K.in
./NPT_simulations/P_1atm/argon_npt_1atm.in
./NPT_simulations/P_10atm/argon_npt_10atm.in

13. Understanding Real Units vs Reduced Units¶

Why We Use Real Units in This Class¶

Real units (units real in LAMMPS):

• ✅ Direct: Temperature = 300 K (not T* = 2.50)

• ✅ Intuitive: Pressure = 1 atm (not P* = 0.002)

• ✅ Comparable: Can directly compare with experimental data

• ✅ Practical: Easy to understand physical meaning

Reduced units (units lj):

• Good for theory and dimensionless comparisons

• Common in older MD papers

• Less intuitive for beginners

Units in LAMMPS Real Mode¶

Argon Parameters (Memorize These!)¶

14. Summary Checklist¶

After completing this tutorial, you should have:

• Created organized folder structure:

  • NVT_simulations/ with 3 temperature folders

  • NPT_simulations/ with 2 pressure folders

• 5 complete LAMMPS input files (.in files)

• Understand the difference between NVT and NPT fixes

• Know how to convert real units (K, atm) to reduced units

File count check:

cd $SCRATCH/lammps_tutorial
find . -name "*.in" | wc -l

Should output: 5

Next Tutorial: Running these simulations on Anvil (Tutorial 8)

15. Troubleshooting¶

Problem: “mkdir: cannot create directory: File exists”

• That folder already exists. Use ls to check what’s there, or use -p flag: mkdir -p foldername

Problem: “No such file or directory”

• You’re in the wrong directory. Use pwd to check location, cd to navigate

Problem: Lost track of where you are

cd $SCRATCH/lammps_tutorial
pwd
ls -R

Problem: Want to start over

cd $SCRATCH
rm -rf lammps_tutorial
# Then start from Section 11, Step 1

16. Class Examples¶

Pre-made complete folder structures are available:

ls /anvil/projects/x-chm250117/class_examples/

If you get stuck, you can copy the complete structure:

cd $SCRATCH
cp -r /anvil/projects/x-chm250117/class_examples/lammps_tutorial ./

17. Further Reading¶

• LAMMPS Manual — Command reference

• Anvil LAMMPS Documentation — Anvil-specific guide

• LAMMPS fix npt — NPT thermostat/barostat documentation

Let’s make sure your input file works before submitting a job.

Step 1: Request Interactive Session¶

srun -p shared -A chm250117 --nodes=1 --ntasks=4 --time=00:10:00 --pty /bin/bash

Wait for the prompt to change (30 seconds to 2 minutes). When you see a new prompt, you’re on a compute node.

Step 2: Load LAMMPS¶

module load gcc/11.2.0 openmpi/4.0.6
module load lammps/20210310

Step 3: Run LAMMPS¶

lmp -in argon.in

You should see output scrolling by:

LAMMPS (10 Mar 2021)
...
Step Temp Press PotEng KinEng TotEng Density
0 4.1700000 16.503422 -2.6604462 6.2487500 3.5883038 0.8000000
100 4.0523344 15.341847 -2.5327691 6.0722148 3.5394457 0.8000000
...

If you see this, success! Your input file works.

Step 4: Exit Interactive Session¶

exit

12. Understanding the Output Files¶

12. Understanding the Output Files¶

After running, check what files were created:

ls -lh

Quick check:

head -20 thermo.out

Look for the temperature column — it should fluctuate around 4.17 (your target).

13. Common LAMMPS Commands Reference¶

14. Exercise: Modify the Temperature¶

Challenge: Create a second input file for Argon at 300 K instead of 500 K.

# Copy the file
cp argon.in argon_300K.in

# Edit it
nano argon_300K.in

velocity        all create 2.50 87287 dist gaussian
fix             1 all nvt temp 2.50 2.50 0.5

dump            1 all custom 500 argon_300K.lammpstrj id type x y z vx vy vz
log             thermo_300K.out

15. Class Examples Directory¶

Pre-made example files are available:

ls /anvil/projects/x-chm250117/class_examples/

Copy an example to your directory:

cp /anvil/projects/x-chm250117/class_examples/argon_example.in ./

16. Troubleshooting¶

Problem: “nano: command not found”

• You’re not on Anvil or in a proper terminal

Problem: “Permission denied”

• Make sure you’re in $SCRATCH, not $HOME

Problem: LAMMPS crashes with “Unknown command”

• Check for typos in command names

• Make sure there are no extra spaces at line beginnings

Problem: “ERROR: Cannot open input script”

• Check filename: ls -lh argon.in

• Make sure you’re in the right directory: pwd

17. Summary Checklist¶

After completing this tutorial, you should have:

• Created argon.in in $SCRATCH/lammps_tutorial/

• Successfully run LAMMPS in an interactive session

• Seen thermo.out and argon.lammpstrj files created

• Understand the 6-section structure of LAMMPS input files

Next Tutorial: Running LAMMPS jobs on the cluster (Tutorial 8)

18. Further Reading¶

• LAMMPS Manual — Searchable command reference

• Anvil LAMMPS Documentation — Anvil-specific setup

• LAMMPS Tutorials — Community tutorials

15. Further Reading¶

• LAMMPS Manual — The official documentation (searchable!)

• Anvil LAMMPS Documentation — Anvil-specific LAMMPS setup and modules

• LAMMPS Tutorial: Argon — Mississippi State tutorial

• Molecular Dynamics with LAMMPS — Community tutorials

• Frenkel & Smit, Chapter 4 — Reduced units and their conversions