Search our client documentation below, optionally filtered by one or more systems.
Supercomputer:
Due to a critical security vulnerability we need to reboot public-facing systems to deploy a mitigation against the vulnerabilit
Search Documentation
Search Documentation
Cardinal
Overview of the High Bandwidth Memory on Cardinal's Dense compute nodes
Cardinal
Compilers
The Cardinal cluster supports C, C++, and Fortran programming languages. The available compiler suites include Intel, oneAPI, and GCC. By default, the Intel development toolchain is loaded. The table below lists the compiler commands and recommended options for compiling serial programs. For more details and best practices, please refer to our compilation guide.
Cardinal
These are the public key fingerprints for Cardinal:
cardinal: ssh_host_rsa_key.pub = 73:f2:07:6c:76:b4:68:49:86:ed:ef:a3:55:90:58:1b
cardinal: ssh_host_ed25519_key.pub = 93:76:68:f0:be:f1:4a:89:30:e2:86:27:1e:64:9c:09
cardinal: ssh_host_ecdsa_key.pub = e0:83:14:8f:d4:c3:c5:6c:c6:b6:0a:f7:df:bc:e9:2e
PyTorch Fully Sharded Data Parallel (FSDP) is used to speed-up model training time by parallelizing training data as well as sharding model parameters, optimizer states, and gradients across multiple pytorch instances. The current version is FSDP2, which is not backwards-compatible with the original FSDP, which is deprecated as of PyTorch 2.11.0+.
Cardinal
The PBS compatibility layer is disabled on Cardinal so PBS batch scripts WON'T work on Cardinal, though it works on the Pitzer cluster. You also need to use
sbatch (instead of qsub) command to submit jobs. Refer to the Slurm migration page to understand how to use Slurm.
Ascend, Cardinal, Pitzer
The default MPI module, MVAPICH, will be updated from version 3.0 to 4.1 on May 12, 2026, during the upcoming scheduled downtime. For more information, please visit:
https://www.osc.edu/resources/technical_support/hpc_changelog/2026/defau...
https://www.osc.edu/resources/technical_support/hpc_changelog/2026/defau...
MVAPICH is a standard library for performing parallel processing using a distributed-memory model.
PyTorch Distributed Data Parallel (DDP) is used to speed-up model training time by parallelizing training data across multiple identical model instances.
Cardinal
AutoDock is a a suite of automated docking tools. It is designed to predict how small molecules, such as substrates or drug candidates, bind to a receptor of known 3D structure. AutoDock has applications in X-ray crystallography, structure-based drug design, lead optimization, etc.
Ascend, Gateway
rclone is a tool that can be used to upload and download files to a cloud storage (like Microsoft OneDrive) from the command line. It's shipped as a standalone binary, but requires some user configuration before using. In this page, we will provide instructions on how to use rclone to upload data from Google Drive. For instructions with other cloud storage, check rclone Online documentation.
Some projects may wish to have a common account to allow for different privileges than their regular user accounts. These are called community accounts, in that they are shared among multiple users, belong to a project, and may be able to submit jobs. Community accounts are accessed using the sudo command.
A community sudo account has the following characteristics:
