The E-HPC client gives job submitters a dedicated interface to submit jobs, track job status, and manage job files—without requiring direct access to the cluster or the E-HPC console.
The E-HPC client is designed for job submitters. Cluster administrators create and manage clusters using the E-HPC console.
User roles
E-HPC distinguishes between two roles:
Cluster administrator: Creates, configures, and scales out E-HPC clusters. Manages job submitter accounts.
Job submitter: Submits jobs and manages job files. Uses the E-HPC client.
Features
| Feature | Description |
|---|---|
| Job query | Query jobs by filter condition to track status and schedule tasks. |
| Job submission | Submit jobs using a guided wizard, a script, or a saved job template. |
| Application Center | Configure and run applications such as GROMACS and Fluent through a visual interface. |
| Data management | Transfer files between your local computer and the cluster using the built-in file browser. |
| Session management | Log in to the cluster remotely via the terminal, or connect to a cloud desktop using Virtual Network Computing (VNC). |
| Operation logs | Query operation logs to audit activity in the E-HPC client. |
Use cases
Choose the deployment model that matches your workflow:
Run the client on a local computer
Use this model when your jobs require no data pre-processing or post-processing. Submit jobs, query status, upload data, and download results directly from your local machine.
Run the client in a cloud desktop
Use this model when your jobs require data pre-processing or post-processing—for example, simulation preparation or results analysis. Install the required pre-processing and post-processing software in a cloud desktop alongside the E-HPC client. This keeps your modeling and analysis results within the same cloud environment.
Connect a cloud desktop to your enterprise Active Directory (AD) domain so that AD users can seamlessly access the cloud environment.
For one-stop cloud HPC solutions that include data pre-processing, data post-processing, parallel computing, and data management capabilities, combine the E-HPC client with Alibaba Cloud Elastic Desktop Service (EDS) and cloud storage.
Limitations
| Limitation | Detail |
|---|---|
| Scheduler requirement | The cluster must use PBS as its scheduler. Other schedulers are not supported by default. To connect to a cluster running a different scheduler, submit a ticket. |
| Concurrent users | A single cluster supports up to 15 simultaneous E-HPC client connections. To raise this limit, submit a ticket. |
| Root access | Logging in to the E-HPC client as a root user is not supported. |
What's next
| Task | Description | Reference |
|---|---|---|
| Deploy an environment | Set up the client environment as a cluster administrator before first use. | Deploy an environment for an E-HPC client |
| Log in | Download the client package, install the client, and log in. | Log on to an E-HPC client |
| Submit jobs | Choose a submission method: wizard, CLI, job template, or Application Center. | Submit a job |
| Run applications | Use Application Center to run GROMACS, High-Performance Linpack (HPL), iPerf, Flexible I/O Tester (FIO), or AutoDock Vina. | Run GROMACS · Run HPL, iPerf, and FIO · Use AutoDock Vina |
| Query job data | View job name, status, start time, end time, and details after submission. | Query job data |
| Manage files | Transfer job files between your local computer and the cluster. Modify, rename, or delete files. | Manage job files |
| Manage sessions | Use the terminal to run cluster commands remotely, or use VNC to manage the cluster through a cloud desktop. | Manage sessions |
| View operation logs | Query operation records by time range and identify exceptions. | View operation logs |