Tuck IT Knowledge Base
Research Computing Hub
High-performance computing, storage, research software, data, AI, and consulting available to Tuck researchers.
Research at Tuck is supported by a substantial ecosystem of high-performance computing, advanced research software, institutional datasets, and expert consulting. This article brings those resources together in one place, with pointers to whoever owns each one.
Who this is for
Tuck faculty and post-doctoral researchers. Tuck Research Computing is primarily faculty-facing. Post-doctoral researchers should route requests through their supervising Tuck faculty member.
Engage early. That matters most for projects involving complex analysis, large datasets, sensitive or regulated data, or anything that will need serious compute. A conversation at the design stage is far cheaper than a rebuild later.
Two Sources of Support
What is listed here comes from two different places, and knowing which is which saves time when something needs attention.
Provider
What they provide
Dartmouth Research Computing and Data
Central, institution-wide
The computing and storage platforms: the Discovery cluster, Andes and Polaris, Open OnDemand, and DartFS. Baseline access is free.
Tuck Research Computing
Part of Tuck IT
Hands-on help: statistical consulting, research design, data extraction and preparation, programming support, and training.
If you are not sure which you need, start with Tuck Research Computing. They will either help directly or point you at the right central service, which is quicker than working it out yourself.
High-Performance Computing
Dartmouth provides a tiered, no-cost research computing environment operated by Research Computing and Data. These systems support compute-intensive, data-heavy, and interactive workflows across disciplines.
Discovery Cluster
For large-scale batch processing and heavy computation.
- Linux-based cluster running RHEL 8
- Roughly 130 compute nodes, over 6,000 CPU cores and 120,000 GPU cores
- More than 12 TB of aggregate memory, and around 3.2 PB of total storage across systems
- Baseline access is free under the Research Computing Bill of Rights
Andes and Polaris
Large-memory and interactive systems.
- Shared-memory systems built for applications that need a lot of RAM
- Suited to GUI-based programs, interactive analysis, and building environments
- Common uses include R package compilation, conda environments, Apptainer images, and software development
- Fast local scratch space makes iterative work noticeably quicker
Open OnDemand
HPC in a browser, with no command line required.
- Web interface for reaching HPC resources
- Run RStudio, JupyterLab, MATLAB, and VS Code directly in the browser
- Full HPC power without needing to log in at a terminal
Research Storage and Cloud Services
DartFS
The research network file system, and the right home for research data.
- 50 GB personal home directory for every researcher
- 1 TB shared lab volume for faculty at no cost
- Mountable from Windows, macOS, and Linux
- Accessible on campus, over VPN, and from the HPC systems
- Automatic daily, weekly, and monthly snapshots
Scratch and secure storage
- Up to 5 TB of scratch space through the Discovery free tier
- Secure lab volumes available for regulated or sensitive data
-
Consult Research Computing early where compliance is a factor. Retrofitting a project for regulated data is considerably harder than planning for it.
Personal cloud storage
- 2 TB OneDrive with enterprise security and daily backup
- Covered by an institutional BAA and approved for all data levels
- Google Workspace storage is available for collaboration
For research datasets, use DartFS or secure HPC storage rather than personal cloud storage. Personal cloud is fine for drafts and working documents. It is not the right place for a dataset that needs to be reproducible, shared with collaborators, or reachable from the cluster.
Statistical, Programming, and Research Software
Category
Available
Core statistical packages
SAS, SPSS, Stata, JMP, MATLAB
Programming and data science
R, RStudio, Python (Anaconda), Julia, Jupyter Notebooks
Mathematical and Bayesian modelling
Mathematica, Maple, JAGS
Qualitative analysis
Atlas.ti, NVivo
Databases and spatial tools
PostgreSQL, GDAL, GEOS
Writing and collaboration
Overleaf Professional, BioRender (discounted)
AI and Generative AI for Research
-
amosAI: centralised access to multiple model families for side-by-side comparison
-
Microsoft Copilot: available through Microsoft 365 with enterprise-grade data protection
-
Dartmouth Chat: commercial and Dartmouth-hosted open-weight models
-
API access: cost-free API access to Dartmouth-hosted local models
Check the data classification before putting research data into any AI tool. Approval differs by tool, and several are not approved for high-risk data. See the Tuck AI Data Sensitivity Guide.
Most of these tools are also consumption-based with monthly limits, and the most capable models carry a high cost per use. If your research involves sustained or large-scale AI work, talk to us before you begin so we can plan capacity. See AI Tools, Access and Consumption Models.
Consulting, Training, and Hands-On Support
Tuck Research Computing works directly with faculty on the analytical and technical side of research projects:
- Statistical consulting, research design, and advanced analysis
- WRDS data extraction, merging, and preparation
- Programming support in Python, R, SAS, Stata, SPSS, MATLAB, and C or C++
- Custom simulations, data transformations, and Excel add-ins
- Workshops, seminars, and one-to-one consultations
How requests are handled
- Services are primarily faculty-facing. Post-doctoral researchers coordinate through their supervising faculty member.
- Work may be billed to faculty accounts. Ask about this when you make the request rather than afterwards.
- Requests are prioritised by research impact and urgency, so lead time genuinely helps.
Getting in Touch
Research Computing works by email rather than through the Tuck IT service portal. If you have already opened a ticket with the Help Desk, mention it and we will connect the two so nothing falls between them.
Related Articles
Tuck School of Business, Dartmouth College. Platform specifications and capacity are set by Dartmouth Research Computing and Data and may change. Last reviewed: October 2026.