Introducing Gaia: The Amsterdam UMC High-Performance Computing Resource

We’re thrilled to announce the official launch of Gaia, our next-generation High-Performance Computing (HPC) platform purpose-built to accelerate scientific discovery, diagnostics, and data-intensive research across our institution.

Whether you’re modeling complex systems, processing large diagnostic datasets, or training large language models, Gaia is is there for you.

What’s Under the Hood?

Gaia can be utilised for research projects as well is diagnostic flows. Depending on you intended use, you can login to Helios.umcinfra.nl (research applications) or Selene.umcinfra.nl (diagnostic applications). Both login nodes can be accessed through SSH or through a graphical web interface based on Open Ondemand.

Compute and GPU Power

From the login nodes you have access to:

  • 9 Standard Compute Nodes
    • 256 CPUs
    • 756 GB RAM
  • 3 GPU Nodes
    • 64 CPUs
    • 756 GB RAM
    • 4x NVIDIA H100 94 GB GPUs
  • 1 Big Memory Node
    • 256 CPUs
    • A massive 2.2 TB RAM for memory-heavy jobs

All worker nodes are interconnected with high speed & low latency HDR Infiniband and optimised for high-throughput, parallel computing tasks.


How to get started?

Existing users are currently being migrated from older systems. As of September 2025 anyone with an Amsterdam UMC account can apply for access to Gaia. If you have any questions or would like to express your interest into obtaining an account, please contact us through ServiceNow (IT Servicedesk, be sure to mention HPC-support). Also, make sure to visit the wiki for more information.


Storage & Data Management

All users have access to the centralized archiving system maintained by our IT department for long-term and secure storage. Whether you need temporary scratch space or permanent archival, Gaia integrates with our storage ecosystem to protect your data throughout its lifecycle.


Cost Model

We strive to provide fair and transparent access. Pricing and billing details can be found here.


Built for Sensitivity and Security

All Gaia resources operate within our internal infrastructure — making it fully suitable for sensitive data handling. Whether you’re dealing with patient data, proprietary models, or unpublished research, Gaia ensures secure compute workflows.


Frequently Asked Questions (FAQ)

Do I need separate accounts to access Helios and Selene

No, you don’t need separate accounts. A single user account provides access to both login nodes and the entire Gaia system.

Are the research and diagnostics systems separate clusters?

No, Gaia is one unified system. Helios and Selene are simply different access points optimized for research and diagnostics workflows, but they connect to the same backend resources.