This guide walks you through a complete E-HPC workflow using GROMACS to simulate water molecules: create a cluster, set up a user, submit a job, and verify the results.
Workflow overview
The following diagram shows the end-to-end workflow.
| Step | Task | Purpose |
|---|---|---|
| 1 | Create a cluster | Provision compute resources for job submission and execution |
| 2 | Create a user | Establish an identity for submitting and managing jobs |
| 3 | Create and submit a job | Run your computation on the cluster |
| 4 | View the performance and result of a job | Verify output and review performance metrics |
Step 1: Create a cluster
Create a cluster to provision the compute resources that your jobs need. A cluster handles job submission, debugging, running, and result access.
For this guide, create a cluster with the following configuration:
| Setting | Value |
|---|---|
| Software | GROMACS |
| Scheduler | PBS |
| Compute node instance type | ecs.gn5-c4g1.xlarge |
After this step, you have a running cluster ready to accept job submissions.
Step 2: Create a user
Create a user to establish your identity in the cluster. The cluster requires a user account to track job ownership and enforce permissions.
After this step, you have a user account that can manage the cluster and submit jobs.
Step 3: Create and submit a job
Create and submit a job to run your computation. A job is the basis of E-HPC. A job consists of a shell script and an executable file. Jobs are queued and dispatched in a sequence determined by specified queues and scheduler.
For this guide, use GROMACS to simulate the motion of water molecules.
After this step, your job is queued and running on the cluster. You can confirm successful submission in the E-HPC console under Jobs.
Step 4: View the performance and result of a job
View the performance and result of a job to verify the output. You can:
Monitor job performance metrics
Review cluster performance history
Profile performance hotspots
Connect to the cluster remotely to inspect the simulation output
After this step, you have completed a full E-HPC workflow from cluster creation to job result verification.
What's next
Scale your cluster by adjusting the number of compute nodes to handle larger workloads.
Submit production jobs by adapting the shell script and executable file pattern from this guide to your own workloads.
Monitor ongoing job performance using the metrics and profiling tools introduced in Step 4.