P1 Compute
Clusters
P1 funds dedicated partitions at three Danish HPC facilities, each matched to a research profile. We help run them as well: the P1 compute coordinator holds admin on the systems and works alongside each provider's HPC team, so the service stays fit for research rather than drifting from what researchers need. The pages below call them clusters, which is how you will experience them day to day; formally each one is a P1 partition on the provider's system. Use the table below to find the one that fits your project, then follow its page for access and usage instructions.
| If you need | Go to | Why |
|---|---|---|
| General ML/AI research on non-sensitive data | DTU HPC | Free compute and storage, batch + interactive H100 access |
| Personal data in a strictly controlled setting | NGC HPC | Air-gapped, GDPR-compliant environment |
| To mature a workload for large-scale runs | Gefion HPC | GDPR-compliant DGX H100 sandbox; the step to full nodes needs its own funding |
Guidance
Unsure which fits, or want a second opinion? The compute coordinator is glad to help: tell us about your workload → and we'll help you find and use the right resource. Entirely optional, and you are just as welcome to head straight to a cluster page.
Open to researchers at any Danish university. Looking beyond P1? See the wider compute ecosystem.
General-purpose HPC at DTU for P1 members, with free compute and storage for medium-scale ML research.
GDPR-compliant HPC at the National Genome Centre for P1 projects handling sensitive or personal data.
A sandbox on Gefion for developing and maturing workloads until they are ready for large-scale runs.