Glossary of Terms

From HPCwiki
Revision as of 14:45, 18 June 2026 by Haars0011 (talk | contribs) (IA migration §9: new Glossary of Terms (P2.22) (via create-page on MediaWiki MCP Server))
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

A short glossary of the terms and commands you come across when using Anunna. Where there is a page with more detail, the entry links to it.

Apptainer
The container runtime used on Anunna (formerly Singularity), which runs software packaged as a self-contained image. See Apptainer.
Bucket
A snapshot of a compiler-and-toolchain generation (for example 2023, 2024) that must be loaded before its modules become visible. See Environment Modules.
Container
A self-contained software image that bundles an application with its dependencies, so it runs the same anywhere. See Apptainer.
GRES
"Generic RESource" — a resource requested with --gres, most often a GPU (for example --gres=gpu:1).
Job array
A single submission that creates many near-identical tasks, each with its own $SLURM_ARRAY_TASK_ID. See Array Jobs.
Login node
The machine you connect to over SSH; an access point for editing, data transfer, and submitting jobs — not for computation. See Login Nodes.
Lustre
The fast, parallel filesystem (mounted at /lustre) used for active compute work. See Compute Storage.
Module
A unit of software loaded with module load, provided through the Lmod environment-modules system. See Environment Modules.
MPI
The Message Passing Interface, used to run a program across many cores or nodes. See Performance Optimization/Multiple nodes (MPI).
NFS
The network filesystem that hosts home directories — reliable but slower than Lustre. See Home Directory.
Open OnDemand (OOD)
The open-source software behind the web portal at apps.anunna.wur.nl. See Portal Overview.
Partition
A named set of compute nodes you submit jobs to (main, gpu, gpu_amd). See Partitions / Queues.
QoS
"Quality of Service" — a policy attached to your account that governs job priority and resource limits.
sacct
SLURM command that reports the accounting record (state, exit code, resource use) of finished jobs. See Monitoring Jobs.
salloc
SLURM command that allocates resources for an interactive session.
sbatch
SLURM command that submits a batch script as a job. See Batch Jobs.
Scratch
Temporary working storage (/lustre/scratch) that is not backed up and may be purged automatically. See Compute Storage.
.sif
The Singularity Image Format file that holds an Apptainer container image. See Apptainer.
sinteractive
Anunna helper that starts an interactive job (a shell on a compute node). See Interactive Jobs.
SLURM
The scheduler that allocates resources and runs jobs on the cluster. See Scheduler Overview (Slurm).
squeue
SLURM command that lists jobs currently known to the scheduler. See Monitoring Jobs.
srun
SLURM command that launches a task, or parallel tasks, within an allocation.
Wall time
The real-world time limit set on a job with --time; the job is stopped when it is reached.

See also