Migrate to Carina 2.0

The original Carina cluster has been replaced with Carina 2.0.  No additional projects will be created on the older cluster.

Carina 2.0 is free to use for any Stanford PI-led team working with high-risk data, with additional paid storage available.

Carina 2.0 runs on five Dell PowerEdge XE7745 compute nodes, each packing 8 NVIDIA L40S GPUs, 256 CPUs/512 threads, 1.5 TB of RAM, and around 28 TB of local NVMe storage. Each project includes 1 TB of storage, with additional storage available in 10 TB increments. Carina 2.0 has Open OnDemand as a web interface, serving up tools like R Studio, VS Code, SAS, Stata, Matlab, and Jupyter Notebooks, right in your browser. 

If you’re still on Carina 1.0, now’s the time to make the move. Team Carina has step-by-step migration guides for both individuals and projects at docs.carina.stanford.edu. 

Reach the team at srcc-support@stanford.edu to plan your migration.

Can I Use Carina 2.0?

The rules for Carina 2.0 are the same as they were for the original Carina.

Research must be Stanford-led

Carina is available to any Stanford researcher who has Principal Investigator (PI) privileges or who is part of a team led by a Stanford PI.

Note: External collaborators, or users without a SUNet ID, can be sponsored by a PI and get a full sponsored account, with SUNet ID and Stanford email. Please see the sponsorship page for more information.

Data must be high-risk

Carina is exclusively available for research using high-risk data.

If your project does not involve high-risk data, Stanford Research Computing has robust research computing options for low- and moderate-risk data.

A Data Risk Assessment must be completed

Each new data set brought into the Carina environment will undergo a Data Risk Assessment. Stanford’s Information Security Office (ISO) and the University Privacy Office (UPO) evaluate projects based on all applicable security and privacy laws and regulations as well as University policy. Once the DRA is approved, the data set can be brought on the Carina platform.

Note: The DRA requirement does not apply if the project only uses data from the Stanford Medicine Research Data Repository (STARR) or Population Health Sciences (PHS), which are already covered by a DRA.