A computer in the HPC cluster is referred to as a node, and there are two types of nodes on Andromeda: login nodes and compute nodes.
Login node:
A login node in the HPC cluster serves as a gateway for users to connect and access all available computing resources. It provides an external interface to the BC computing cluster. As shared resources, login nodes are not intended for executing computing tasks; instead, they are designed for tasks such as submitting jobs and handling activities related to job preparation, monitoring, and data management. Please note, login nodes must not be used to execute computing jobs..
Compute node:
Compute nodes are the machines responsible for executing the actual computations and calculations requested by users. On the Andromeda cluster at Boston College, Slurm (Simple Linux Utility for Resource Management) serves as an open-source job scheduler and workload manager in HPC environments. The job scheduler allocates resources on compute nodes to execute these tasks in parallel, enabling efficient processing of large-scale computations. By default, users are not allowed to log in to compute nodes unless they have a job running there.