Use GROMACS to perform a molecular dynamics simulation

Background

GROMACS (GROningen MAchine for Chemical Simulations) is a software package for molecular dynamics simulation. It models Newtonian equations of motion for systems with millions of particles, making it suited for biochemical research involving proteins, lipids, and nucleic acids. GROMACS supports common MD algorithms and can use GPUs to accelerate computing-intensive tasks. For more information, see the GROMACS official website.

Sample datasets

Two datasets are available for this tutorial. This tutorial uses Example 2.

• Example 1: Lysozyme in water — simulates a protein interacting with ions in a box of water. Download: https://public-ehpc-package.oss-cn-hangzhou.aliyuncs.com/Lysozyme.tar.gz

• Example 2: Motion of water molecules — tracks the motion of water molecules in a specific space at a specified temperature. Download: https://public-ehpc-package.oss-cn-hangzhou.aliyuncs.com/water_GMX50_bare.tar.gz

Prerequisites

Before you begin, ensure that you have:

• An E-HPC cluster with the following configuration: For cluster creation instructions, see Create a cluster by using the wizard.

  Setting	Value
  Deploy mode	Standard
  Nodes	2 management nodes, 1 compute node, 1 logon node
  Compute instance type	GPU-accelerated (e.g., ecs.gn5-c8g1.2xlarge)
  Image	CentOS 7.2 (public)
  Scheduler	PBS
  VNC	Enabled

• A cluster user with the following settings: This user logs on to the cluster, compiles software, and submits jobs. For instructions, see Create a user.

  Setting	Value
  Username	gmx.test
  User group	sudo permission group

• The following software installed on the cluster. This tutorial was tested with these versions — other versions may behave differently. For installation instructions, see Install software.

  Software	Version
  Gromacs-gpu	2018.1
  openmpi	3.0.0
  cuda-toolkit	9.0
  VMD (Visual Molecular Dynamics)	1.9.3

Step 1: Connect to the cluster

Connect using either the E-HPC client or the E-HPC console. After connecting, you are logged on to /home/gmx.test.

Option A: E-HPC client

1. Start and log on to the E-HPC client.

2. In the left-side navigation pane, click Session Management.

3. In the upper-right corner of the Session Management page, click terminal to open the Terminal window.

Option B: E-HPC console

1. Log on to the E-HPC console.

2. In the upper-left corner of the top navigation bar, select a region.

3. In the left-side navigation pane, click Cluster.

4. On the Cluster page, find the cluster and click Connect.

5. In the Connect panel, enter the username and password, and click Connect via SSH.

Step 2: Submit a job

1. Download and decompress the sample dataset.

   wget https://public-ehpc-package.oss-cn-hangzhou.aliyuncs.com/water_GMX50_bare.tar.gz
   tar xzvf water_GMX50_bare.tar.gz
   chown -R gmx.test water-cut1.0_GMX50_bare
   chgrp -R users water-cut1.0_GMX50_bare

2. Create a job script named gmx.pbs.

   vim gmx.pbs

   Paste the following content into the file:

   #!/bin/sh
   #PBS -j oe
   #PBS -l select=1:ncpus=8:mpiprocs=4
   #PBS -q workq
   
   # Load module environment
   export MODULEPATH=/opt/ehpcmodulefiles/
   module load gromacs-gpu/2018.1
   module load openmpi/3.0.0
   module load cuda-toolkit/9.0
   export OMP_NUM_THREADS=1
   
   cd /home/gmx.test/water-cut1.0_GMX50_bare/0096
   
   # Preprocess input files to generate a TPR binary input file
   /opt/gromacs-gpu/2018.1/bin/gmx_mpi grompp -f pme.mdp -c conf.gro -p topol.top -o topol_pme.tpr
   
   # Run the simulation: 4 MPI processes, 1 OpenMP thread each, 100,000 steps
   mpirun -np 4 /opt/gromacs-gpu/2018.1/bin/gmx_mpi mdrun -ntomp 1 -nsteps 100000 -pin on -s topol_pme.tpr

   Key parameters:

   Parameter	Value	Purpose
   ncpus=8	8 vCPUs	Matches the ecs.gn5-c8g1.2xlarge instance (8 vCPUs, 1 P100 GPU)
   mpiprocs=4	4 MPI processes	Splits the simulation across 4 parallel processes
   -ntomp 1	1 OpenMP thread per MPI process	Avoids over-subscribing CPU cores — 4 processes x 1 thread = 4 threads on 8 vCPUs
   -nsteps 100000	100,000 iterations	Controls simulation length; increase for longer or more accurate runs
   -pin on	CPU affinity enabled	Pins each thread to a fixed CPU core, reducing context-switching overhead

   Adjust ncpus, mpiprocs, and -nsteps based on your computing requirements.

3. Submit the job.

   qsub gmx.pbs

   The output is the job ID, for example:

   0.scheduler

Step 3: View the job result

1. Check the job status.

   qstat -x 0.scheduler

   Expected output:

   Job id            Name             User              Time Use S Queue
   ----------------  ---------------- ----------------  -------- - -----
   0.scheduler       gmx.pbs          gmx.test          00:34:42 F workq

   In the S column, R means the job is running and F means it has finished. While the job is running, monitor the live output with:

   tail -f gmx.pbs.o0

   After the job finishes, view the full output with:

   cat gmx.pbs.o0

2. Visualize the result using VNC and VMD.

   Enable VNC:

   Note

   Make sure the VNC ports are open in the security group for the cluster. The console automatically opens port 12016. If using the client, open the ports manually. Port 12017 supports one user at a time — open additional ports starting from 12017 for each additional user.

   • Using the client: In the left-side navigation pane, click Session Management. In the upper-right corner, click VNC to open VNC Viewer.

   • Using the console:

     1. In the E-HPC console, click Cluster in the left-side navigation pane.

     2. Select the cluster. Choose More > VNC.

     3. Connect to the visualization service. For details, see Use VNC to manage a visualization service.

   Open and load the result in VMD:

   1. In the cloud desktop, choose Application > System Tools > Terminal.

   2. Run /opt/vmd/1.9.3/vmd to open VMD (Visual Molecular Dynamics).

   3. On the VMD Main page, choose File > New Molecule....

   4. Click Browse... and select conf.gro.

      Note

      The file path is /home/gmx.test/water-cut1.0_GMX50_bare/0096/conf.gro.

   5. Click Load. The simulation result appears in the VMD 1.9.3 OpenGL Display window.

   

Step 4: View computing performance

1. Log on to the E-HPC console.

2. In the left-side navigation pane, choose Job and Performance Management > E-HPC Tune.

3. Find the cluster and click Node.

View node performance (NodeView tab):

1. Select a job and a compute node.

2. (Optional) Set a time period. When you select a job, the time period automatically matches the job's run window. Adjust it manually as needed.

3. Click Metrics and select the metrics to view.

4. Hover over the graph to see detailed data.

View process performance (ProcView tab):

1. Click the ProcView tab.

2. Select a job and a compute node.

3. (Optional) Set a time period.

4. Hover over the graph to see detailed data.

Start a performance profiling task:

1. Click a point in time on the process performance graph, then click the process to profile.

   

2. In the dialog box, set the profiling duration and frequency, and click OK to start real-time profiling.

View profiling results (Profiler tab):

1. Click the Profiler tab.

2. Select a profiling task and click View.

3. Hover over the graph to see detailed data.

Related topics

Use GROMACS to perform a molecular dynamics simulation