Research Computing (RC) provides a growing shared high performance computing environment known as the O2 Cluster, which accommodates diverse requirements and workflows for HMS-affiliated researchers. Tens of thousands of jobs run on O2 every day for big and small projects in next-gen sequencing analysis, molecular dynamics, mathematical modeling, image analysis, proteomics, and other areas. Research Computing is constantly working to improve O2’s job handling software and configuration to balance throughput between our many users. 

Request an O2 account with the Account Request form

For more information, please view our O2 documentation

----------------

CPU Processing power: The O2 high performance computing cluster of over 11,000 compute cores is dedicated to research activities. It includes 350 nodes, each with 256 GB RAM and 32 compute cores and 9 high-memory nodes with up to 1 TiB RAM. For more information, visit the Research Computing CPU webpage.

GPU Processing power: Demand for GPU computation is growing rapidly. Through 2020, O2 had 8 V100 GPU cards and 24 older double-precision cards. In 2021, Research Computing has added 47 RTX8000 single-precision and 24 V100S double-precision GPU cards. For more information, visit the Research Computing webpage on GPU resources.

Storage: HMS IT provides different types of storage for different purposes and stages of the research lifecycle. Active Compute storage is a 10 PiB Isilon all-flash system, and Scratch is a 1 PiB all-flash storage system. Active Collaboration provides desktop access to 16 PiB of Isilon spinning disk storage. Standby provides 6 PiB on slightly less performant Isilon spinning disk. For more information, go to our Storage page.

Scientific Applications: The O2 cluster has 170+ applications and tools for computational analysis. For more information, please see our full list of Currently Available Software on O2.

* O2 Portal: Web-based interface that provides easy access to the O2 cluster. Includes tools for viewing/editing files, managing cluster jobs, and accessing login shells. Also supports interactive graphical programs like R Studio, Jupyter notebook, MATLAB. Visit the O2 Portal webpage.

User Training

Courses include an introduction to the O2 HPC cluster, Python, R/Bioconductor, MATLAB, and more. Training classes are open to all HMS affiliates. Further information about sessions and registration can be found on the Research Computing Trainings webpage. 

Contact: rchelp@hms.harvard.edu

Research Computing Office Hours are held every Wednesday from 1-3pm.