Overview

Login nodes are shared gateways to SeaWulf. They are meant for light tasks such as editing files, submitting jobs, compiling code, and monitoring job progress. Heavy computation should always be run on compute nodes.

Appropriate Usage

• Edit scripts, view files, and manage data.
• Compile code and load modules.
• Submit and monitor SLURM jobs (sbatch, squeue, etc.).
• Perform lightweight testing or small-scale calculations.

What to Avoid

• Running long or resource-intensive computations directly on the login nodes.
• Starting jobs that use multiple cores, GPUs, or large amounts of memory on login nodes.
• Leaving idle processes that consume memory or CPU unnecessarily.
• Large file transfers that can saturate the network.

Tips for Responsible Usage

• Use ssh with caution and log out when finished.
• Use srun --pty bash or SLURM batch scripts to run interactive or large jobs on compute nodes.
• Monitor login node usage and avoid exceeding CPU or memory limits.
• Contact support via iacs.supportsystem.com if you need help with job placement or node selection.