The VU MIF High Performance Computing (HPC) system provides access to powerful computing resources for tasks that are too large, too slow, or require more memory than a typical personal computer can provide. HPC can be used for scientific computing, simulations, data analysis, artificial intelligence, machine learning, and other research workloads.
Users connect to the system remotely, store their data in a shared file system, and submit jobs through the SLURM scheduler. The scheduler automatically allocates the required resources and runs computations on available compute nodes.
New users should complete the following tutorials in order:
| Problem | Possible Cause | Solution |
|---|---|---|
| Cannot connect to the HPC system via SSH | Incorrect hostname, username, VPN requirement, or network issue | Verify the connection settings and ensure you are using the correct username. See Connection. |
| “Permission denied” when connecting | Incorrect password, SSH key configuration, or missing account access | Verify your credentials and ensure your HPC account has been activated. |
| Command not found | Software module is not loaded or command is misspelled | Check the command syntax and load the required software module. |
| Job remains in PENDING state | Requested resources are not currently available | Use squeue to check the reason and wait for resources to become available. |
| Job finished but no output was produced | Application error or output written to a different location | Review job output files and verify file paths. |
| Job was cancelled or failed | Application error, exceeded memory limit, or exceeded runtime limit | Check the SLURM output and error files for details. |
| “Disk quota exceeded” or “No space left on device” | Storage quota has been reached | Remove unnecessary files or move data to another storage location. |
| File cannot be found | Incorrect path or directory | Use pwd and ls to verify your current location and available files. |
| Unsure where to start | New to Linux or HPC environments | Follow the tutorials in the recommended order below. |
If you encounter a problem that is not covered in this guide:
When reporting an issue, providing the exact error message and job ID usually allows problems to be resolved much faster.