The Tufts HPC Cluster
The Tufts HPC Cluster is Tufts University's primary research platform for large-scale computing and AI research, available at no cost to all faculty, researchers, students, and staff. Whether supporting groundbreaking research, graduate studies, or coursework, the platform is designed to meet foundational computational needs of the Tufts community for teaching, learning, and research.
A Capable Research Platform
The Tufts HPC Cluster provides the Tufts community access to some of today's most powerful computing hardware. With over 200 nodes spanning thousands of CPU cores and hundreds of GPUs, including cutting-edge NVIDIA H200 GPUs, the platform is built to handle the full range of modern computational workloads including AI/ML, bioinformatics, simulations, statistical analysis, and data-intensive research more broadly. Nodes with up to 1.5TB of RAM and high-core-count CPUs ensure that even the most memory-intensive workloads have the resources to run efficiently. In addition, a100Gbps network enables fast, reliable communication across the cluster and beyond. While the scale of platform meets many needs, some computationally intense research groups invest in additional dedicated cluster “contrib” nodes.
The cluster features a multi-petabyte all-flash parallel filesystem designed to scale with the demands of modern research. The storage system is organized into three tiers that provides users the flexibility to balance performance, capacity, and cost based on their needs. Data is accessible from any node on the cluster, and large-scale transfers are supported through the Globus file transfer service, making it easy to move data to and from collaborators and external resources.
Convenient to Use
While researchers can use a traditional command line to access the Tufts HPC Cluster, additional options are available. The OnDemand web portal provides browser-based access to the cluster, allowing users to launch jobs, manage files, monitor active work, and check real-time cluster status, all from a simple interface. Web applications walk users through submitting jobs for common applications, and many tools are available to run directly in the browser. Dedicated TTS Research Technology support staff are available to help users at every level, from first-time users getting oriented to experienced researchers tackling complex workflows. Extensive documentation, available on the RT Guides website, ensures that help is always within reach.
A Rich Software Environment
The Tufts HPC Cluster comes with a broad library of pre-installed software, accessible through a straightforward modules system that requires no manual installation. Researchers can get to work immediately with popular tools including RStudio, Jupyter Notebooks, MATLAB and VSCode. The library includes applications common to AI, bioinformatics, computational chemistry, geospatial, digital humanities, statistics, physics, and engineering simulations. When specialized tools are needed, TTS Research Technology staff can assist with sourcing and installing both third-party and custom software packages, ensuring researchers always have the tools their work demands.
A Platform That Grows with You
The Tufts HPC Cluster operates under a model that benefits both individual researchers and the broader Tufts community. In addition to the university’s shared public resources, individual researchers and departments requiring significant dedicated compute may purchase their own compute nodes, known as “contrib” nodes. TTS Research Technology manages system administration so compute node owners have no infrastructure management responsibilities. Owners are responsible for covering the purchase price and minimal initial installation costs. Owners receive priority access to their “contrib” nodes and maintain control over access policies. When those nodes are not in use, their capacity is returned to the shared public pool, expanding the resources available to the entire Tufts community. This collaborative model ensures that investment in the platform benefits everyone. Tufts faculty should reach out to TTS Research technology if they are considering purchasing their own compute nodes.
Connected to a Regional Research Community
The Tufts HPC Cluster is housed at the Massachusetts Green High Performance Computing Center (MGHPCC), a state-of-the-art intercollegiate facility shared by leading research universities in the region. The MGHPCC supports a community of more than 20,000 faculty, students, researchers, and educators, all backed by dedicated on-campus research teams. Being part of this broader ecosystem means Tufts researchers benefit from a facility built and operated to the highest standards in the industry.