Getting Started¶
Directory Structure¶
When you log into the head node, you will find the following shared directories. These are stored on the NAS and accessible from all nodes.
| Path | What it is | Example |
|---|---|---|
/tml/<username> |
Your home directory. Store your project files and simulation data here. | /tml/shlee |
/nsource |
Software installation sources (ANSYS ISOs, drivers, etc.). Read-only for most users. | /nsource/ansys |
/nopt |
Shared software installations and scripts used cluster-wide. | /nopt/ansys_inc, /nopt/common |
/nas.12f |
NAS at SNU Bldg 301, 12F lab. Currently not mounted. | |
/nas.14f |
NAS at SNU Bldg 301, 14F lab. Mounted on head node only. |
Your home directory (/tml/<username>) is where you should do your work. It is shared across all nodes, so files you save here are available when your job runs on any computing node.
User Account¶
To use the cluster, you need an account created by the admin.
- All accounts are managed by a central LDAP server. A single account gives you access to all nodes and NAS.
- Your username will typically be a short form of your name (e.g.
shlee) - Your home directory will be created at
/tml/<username> - You will be added to the SLURM scheduler so you can submit jobs
- Contact the cluster admin to request an account
Rules¶
- Do not run simulations on the head node. The head node is shared by all users for login and file management. Running heavy computations here will slow it down for everyone. Always submit jobs through SLURM.
- Return software licenses after use. Commercial software has a limited number of licenses. Close the application when you're done so others can use it. See the Software Installation section for details.
- Do not touch anything outside your own folder. Only modify files in your home directory (
/tml/<username>). Do not modify or delete files in other users' directories,/nsource,/nopt, or any shared location. - Manage your storage and keep backups. The NAS is shared by everyone. Clean up unnecessary files regularly and keep your own backups of important data — data loss can happen, and there is no guarantee of recovery.
- Report issues to the admin. If a node is down, a job is stuck, or something looks wrong, let the admin know.