Main Page: Difference between revisions
mNo edit summary |
No edit summary |
||
| (3 intermediate revisions by 2 users not shown) | |||
| Line 115: | Line 115: | ||
* <h3>[[Debugging Jobs]]</h3> Working out why a job failed | * <h3>[[Debugging Jobs]]</h3> Working out why a job failed | ||
* <h3>[[Checkpointing]]</h3> Saving and restarting long jobs | * <h3>[[Checkpointing]]</h3> Saving and restarting long jobs | ||
* <h3>[[Scheduled tasks (cron)]]</h3> Running recurring tasks with scrontab | * <h3>[[Scheduled tasks (cron)|Scheduled Tasks (cron)]]</h3> Running recurring tasks with scrontab | ||
* <h3>Parallel Workflows</h3> | * <h3>Parallel Workflows</h3> | ||
| Line 123: | Line 123: | ||
** <h4>[[Workflows/Multi-threaded|Multi-threaded]]</h4> Many cores on one machine, within a single process | ** <h4>[[Workflows/Multi-threaded|Multi-threaded]]</h4> Many cores on one machine, within a single process | ||
** <h4>[[Workflows/Multi-Process|Multi-Process]]</h4> One calculation spread across several machines | ** <h4>[[Workflows/Multi-Process|Multi-Process]]</h4> One calculation spread across several machines | ||
* <h3>[[Agentic AI skill]]</h4> How to instruct your AI agent to use Anunna | |||
== Quick links == | == Quick links == | ||
Latest revision as of 12:13, 5 November 2026
Anunna is a High Performance Computing (HPC) cluster hosted by Wageningen University & Research. It is open to all WUR research groups, and to other organisations and companies running collaborative projects with WUR.
You reach the cluster over SSH at login.anunna.wur.nl, or through your browser at the Apps Portal (https://apps.anunna.wur.nl/).
- What Anunna is for and who runs it
- How we decide what to work on
- How the cluster is put together
- The node types and their hardware
- The storage tiers at a glance
- Costs of using Anunna
- Interconnect and data-security posture
- How Anunna came to be
- Green-HPC policy and measures
- Frequently Asked Questions
- Who may use Anunna and on what terms
- Eligibility
- How to request an account
- What is expected of you as a user
- Logging in over SSH to the cluster
- Using applications on Anunna through your web browser
- Moving your work to the cluster from your computer
- Guidance and standard text that can be used in grant proposals, research papers, theses and other publications.
- The entry point you connect to
- Where your jobs run, and how to reach them
- The storage tiers
- Storage limits and how to check them
- Reaching Anunna from off campus
- Moving data to and from Anunna
- How scheduling works
- The available partitions
- Targeting particular hardware
- Submitting a batch script
- An interactive shell on a compute node
- Running many similar jobs at once
- Checking on your jobs
- Stopping a job
- Reserving nodes
- Using shared resources considerately
- The software landscape
- Loading software with modules and buckets
- Installing into your own space
- Software that needs a licence
Scripting languages
- Python modules and environment managers
- R modules, packages, and parallel R jobs
- Loading Julia, packages, and running it in a job
Containers
- Portable, reproducible containers for HPC
MPI implementations
- The MPI library Anunna is built around
- Intel MPI, with the Intel compilers and MKL
- The storage tiers at a glance
- Your personal space
- The fast Lustre filesystem for active work
- Sharing data within a group
- What is backed up, and what is not
- Long-term storage on tape
- Storage limits
- How data moves from active to archived to removed
- Keeping data safe, tidy, and cheap
- Moving data efficiently and reliably
- Confidential and personal data
Graphical Interface Applications
- What the portal is and how the dashboard is laid out
- The general way to launch and connect to a graphical application
- Start a full Linux desktop in your browser
- Jupyter notebooks
- The RStudio IDE for R and Python
- Manage your files in the browser
- A terminal in the browser
- See your jobs in the browser
- Course slides and self-study resources
- Hands-on tutorials
- Instructor-led courses and their dates
- Who to contact and how
- The WUR support portal
- Writing a good problem report
- Current problems and workarounds
- The Anunna user community
- Planned downtimes
- Good and bad practice at a glance
- Working with groups
- Requesting more storage
- Requesting larger or dedicated allocations
- Giving access to collaborators outside WUR
- Tracking your group's usage and costs
- Facility descriptions and support for funding applications
- Moving your work to the cluster
- Keeping your work reproducible
- Managing multi-step pipelines
- Working out why a job failed
- Saving and restarting long jobs
- Running recurring tasks with scrontab
Parallel Workflows
- What parallel computing means on Anunna, and which type fits your work
- One program on a single core — the baseline
- Running the same program many times over
- Many cores on one machine, within a single process
- One calculation spread across several machines
- How to instruct your AI agent to use Anunna
Quick links
- How to Get Help — contact the HPC team
- Maintenance Schedule — planned downtimes
- Glossary of Terms — common HPC and Anunna terms explained