Skip to content

Smithsonian High Performance Computing

Hydra

The SI HPC cluster contains 74 compute nodes, including GPU nodes, with 5,840 CPU cores and 49 TB of memory. The system is designed and operated by the Data Platforms and Advanced Computing team in the Office of Digital and Innovation.

  • Getting Started

    Request an account, log in, submit your first job.

  • Running Jobs

    Queues, job scripts, memory and CPU requests.

  • Interactive Use

    Jupyter, RStudio, VS Code, interactive shells on compute nodes.

  • Storage

    Which filesystem to use, quotas, snapshots, the scrubber.

  • Software

    Modules, compilers, Python and R, bioinformatics packages, and CUDA for GPUs.

  • Data Transfer

    Moving data on and off the cluster with scp, rsync, Globus, or rclone.

Help

News

December 16, 2025. The production RStudio server was upgraded with an updated OS and R version.

December 8, 2025. QSubGen now includes virtual memory and GPU options.

December 5, 2025. Scrubbing of files older than 180 days on the public disks resumes Sunday Dec 7; it had been suspended before the move to Ashburn.