Obtaining an Account
Accounts are tied to allocations controlled by Principle Investigators (PIs). You must first receive an allocation before you can begin using UVA's high-performance computing (HPC) system. All Virginia faculty (including postdoctoral associates) and staff with a valid UVA computing ID can apply for an allocation. Students who wish to use the system must be added to a MyGroups account by an authorized faculty or staff sponsor. Accounts are created automatically but it can take up to an hour for the system to update to recognize the new account.
Please review the guide for new users for information about basic system usage before logging on for the first time.
New users are invited to attend one of our free one-hour orientation sessions ("Introduction to the HPC System") held on Fridays at 1023 Millmont Street, North Grounds at 4 p.m. or by appointment. We can also come to your department or organization. Please use our contact form to reserve your space.
ARCS offers various training opportunities throughout the year to those who need additional help using the HPC system.
Access to the HPC cluster requires a valid Eservices password. Your Netbadge password is not necessarily the same thing, so if you are unable to log in, you should first try resetting your Eservices password here. If the problem persists, contact ITS (which manages all Eservices accounts) through the group's online help desk.
Login and File Transfer Tools
Several options are available for logging in, depending on the operating system you use to access the system. You may also wish to install tools to transfer files to and from your workstation.
A "desktop" environment is available through FastX.
For large file transfers, we recommend Globus.
Each user has a home directory and a scratch directory. The "/scratch" directory is not intended as permanent storage and is not backed up, but is fast and has a large per-user quota, so is the recommended working directory for active jobs. The user /scratch directory should be used for all parallel jobs and for all "high throughput" workflows (many serial jobs).
If you are unfamiliar with Unix we have several short tutorials online.
Dozens of software packages are available pre-installed, Groups may also install their own software as long as they are legally licensed or it is open source.
Software is accessed through environmental modules.
Production jobs must be submitted to the backend compute nodes through the SLURM queueing system.
ARCS staff can assist you with code optimization and parallelization, effective use of Rivanna, and code development, both for HPC and other targets. If you are interested in a collaboration, visit our consulting page.