If you are writing a grant proposal for research that will use CCV facilities, please use the following text as a short description of our facilities:
CCV provides high-performance computing (HPC) resources and scientific computing expertise to Brown University’s research community. Oscar, CCV’s primary research computing cluster, has
- 2 login nodes
- 405 diskless multi-core nodes
- 11552 CPU cores
- available memory ranging from 93 GB to 2096 GB per node
- sharing a high-performance interconnect and file system and running Red Hat Enterprise Linux 7.3
- 27 GPU nodes
- total of 193 GPUs
- available memory ranging from 11 to 32 GB VRAM per GPU
- xCAT (IBM’s cluster administration toolkit) for provisioning all nodes
- a GPFS parallel filesystem which provides a total of 3 PB of usable storage space
- 100 Gb/s EDR Infiniband connectivity
- For more technical details, please see this link.
CCV also supports the Carney condo for The Robert J. & Nancy D. Carney Institute for Brain Science. The condo has
- 10 Intel Cascade Lake nodes, each with 32 cores and 376GB memory
- 100 Nvidia Quadrortx GPUs, each with 24GB memory
- 1.4 TB data storage
Use of dbGaP data on the computing cluster is restricted to a Singularity software container mechanism that isolates the application from outside networking. It also restricts storage write access to a special area on the enterprise Isilon platform that is reserved for dbGaP use. A separate transfer gateway system with file access auditing provides a path for dbGaP data transfer into the system, as well as the transfer of computational results out of the system.
Stronghold is a secure computing and storage environment that enables Brown researchers to analyze sensitive data while complying with regulatory or contractual requirements. It has been established by Brown University to be compatible with Federal and Rhode Island Law standards for data privacy and protection. Each Principal Investigator (PI) is given a dedicated environment for their project to support their researchers, graduate students, and collaborators.
CCV’s state-of-the-art infrastructure is maintained and operated by CCV staff, who have extensive experience in operating shared computational clusters. CCV staff members are responsible for scheduled maintenance, access control, and integration with research-specific hardware as required by researchers. A large collection of software is available on Oscar, including: R, Matlab, Mathematica, Maple, and a large range of optimized math and science libraries and domain-specific applications. CCV staff members are available to assist with the acquisition and installation of any applications not already available on Oscar. CCV staff members are also available to consult on research projects and work directly to support and develop the advanced computational methods used by researchers at Brown. Please contact us if you need a more detailed description.