Research Computing refers to computational data capture and analysis that cannot be trivially performed on a standard desktop or laptop computer. The Kennedy Institute routinely generates and analyses significant quantities of genomic, single-cell and imaging data which must be housed and processed on high-performance compute and large-scale and very fast storage platforms.
We work closely with the BMRC to access compute and storage capabilities at scale. This strategic collaboration enables our researchers to access over 7000 compute cores, 60 GPGPU cards and nearly 20 petabytes (equivalent to around 20 million desktop computer drives) with the ability to run the latest software to support new approaches to data analysis. On top of this, a dedicated RStudio server and Jupyter notebook capabilities support ad hoc exploration and analysis of data as well the development and maintenance of standard analysis pipelines. High speed network connections between the Institute and the BMRC ensure rapid data access.
In addition, we maintain a significant virtual server and storage platform to provide dynamic server provisioning capabilities to support research informatics needs [link to this], web applications, databases and infrastructure monitoring. This allows us to rapidly support researchers who need to explore new software technologies and analytics which may not be suitable for the high-performance compute and storage BMRC platform.
The use of research computing is no longer the domain of computational experts with all research groups within the Institute routinely using the platform to support their research. The platform is accessible to all members of the Institute and is supported by the Research Computing Administrator who provides direct hands-on support for all users, providing expertise in software provisioning, training and data management.