Proposal Resource Kit

Research Computing & Informatics (RCI). Stony Brook University provides centrally managed research cyberinfrastructure through RCI, a unit of the Division of Information Technology. High-performance computing is available on four clusters: SeaWulf (400+ nodes, ~23,000 cores, ~1.86 PFLOP/s, with Intel Sapphire Rapids HBM and NVIDIA GPU nodes), NVwulf (24 NVIDIA H200 NVL GPUs, up to 80 PFLOPS FP8 for AI/ML), Ookami (176 Fujitsu A64FX nodes), and ClinWulf, a HIPAA-compliant cluster (30 nodes, 32 GPUs incl. 8× A100, high-memory nodes to 1 TB RAM, 1.5+ PB VAST storage) for protected health information. Researchers also have access to virtual machines, a research Azure cloud environment, centrally managed academic and clinical storage, and Box for collaboration. Informatics platforms include REDCap, TriNetX, OnCore, and honest-broker and de-identification services. RCI staff provide consultation, training, and grant support.

Stony Brook Medicine research systems that handle protected health information operate under HIPAA and, where applicable, NIST SP 800-171 controls. Clinical data access is governed by IRB approval and CITI training; the ClinWulf environment and clinical research storage provide secure, access-controlled, auditable handling of PHI. De-identification review and honest-broker services are available for compliant data sharing.

Data generated by this project will be managed in accordance with the [NIH/NSF] data management and sharing policy. Active research data will be stored on Stony Brook's centrally managed research storage (backed up where required) or, for protected data, on HIPAA-compliant clinical storage. Data will be documented with appropriate metadata and, at the end of the project, shared through [an appropriate discipline repository / a generalist repository], with de-identification review performed by RCI's honest-broker service where human-subjects data are involved. A data management plan will be prepared using the DMPTool.