High-Performance Computing
Three advanced HPC clusters (SeaWulf, NVwulf, and ClinWulf) for computational research at every scale, with secure, compliant environments for sensitive data: HIPAA-compliant computing for PHI and support for FERPA-protected student records.
Our clusters
Pick the right cluster for your work
SeaWulf
General-purpose HPC: CPU and MPI jobs, large parallel runs
- High-performance cluster with components from AMD, Dell, HPE, and others
- Housed at the Stony Brook University Computing Center
- Available to the campus community and Brookhaven Lab staff
NVwulf
AI and machine learning: GPU training and inference
- Newest GPU-accelerated cluster for AI and machine learning research
- 24 NVIDIA H200 NVL GPUs
- Up to 80 petaFLOPS FP8 performance
ClinWulf
Clinical research with PHI (HIPAA-compliant)
- 30 high-performance nodes, 32 GPUs including 8 NVIDIA A100s
- High-memory nodes up to 1,000 GB RAM
- 1.5+ PB VAST storage
SecureWulf
Secure computing for controlled-access and regulated research data
- TiCrypt secure enclave aligned with NIST SP 800-171: encrypted, isolated project spaces
- 8 GPUs (NVIDIA H200 NVL and RTX PRO 6000 Blackwell) with Slurm
- For approved projects, such as dbGaP or All of Us controlled-tier data
Planning a project that will need it? Tell us now.
Get started
From zero to your first job in three steps
Compute from your browser
SeaWulf's Open OnDemand portal launches applications directly in a web browser. No SSH required.
Help and resources
Everything else you might need
You may also need
Cloud Computing
Research-specific Microsoft Azure environments to host data-intensive projects, run complex simulations, and access powerful computing without on-premises constraints.
Learn more →Virtual Machine
A dedicated environment for researchers to prototype code and applications that cannot run on HPC. Scalable, managed, on-premises VMware infrastructure in Linux and Windows.
Learn more →Academic Research Storage
Centrally managed storage for active research projects with unrestricted data, up to 10 TB per research group on request.
Learn more →