Policies and Guidelines

Policies and Guidelines#

SMU Policies#

Users must comply with all applicable laws, regulations, and SMU policies when using HPC resources. This includes, but is not limited to, SMU’s acceptable use policy

O’Donnell Data Science and Research Computing Institute Community Guidelines#

Please review the ODSRCI Guidelines

Data Restrictions#

Please review SMU’s Institutional Data Governance .

SMU’s HPC systems are generally not approved for use with restricted or confidential data. This includes, but is not limited to:

  • personally identifiable data

  • protected health information

  • FERPA protected data

In most cases, unpublished research is classified as private data according to SMU policies. Private data can generally be used on the HPC systems. However, SMU’s HPC system is a research and academic tool and may lack adequate protections for some types of private data.

Data classified as public can always be used on the system as long as owning, storing, or using the data does not violate other SMU policies, contractual agreements, laws, or other regulations.

Please consult with SMU’s Office of Research and Innovation (ORI) to ensure compliance before utilizing HPC resources.

Purge Policy#

Data stored in $SCRATCH on M3 or the SuperPod has a time based quota. Data older than 60 days will be purged (deleted) without warning. $SCRATCH is intended as high performance, temporary storage for jobs.

If data is needed for a longer period of time, it should be stored in ColdFront storage allocations (limits vary and require justification) or $HOME directories (200 GB). If those are also insufficient, please contact us to discuss options.

Important

SMU does not currently have facilities for archival storage of large datasets. Most of our HPC storage is redundant, but it is not (and cannot) be backed up. Storage space is also limited and current and active usage is prioritized. Please contact us to discuss needs and potential options.

Login Nodes#

Running code, tasks, and persitent processes on login nodes is forbidden. The login nodes are a shared environment and running intensive tasks can make the entire system unstable for all users. Tasks running on login nodes may be terminated by system adminstrators to preserve system stability.

Some examples of acceptable use of login nodes:

  • Submitting and monitoring jobs.

  • Editing or viewing files with text editors, IDEs, or similar.

  • Compiling code (compilation tasks that are expect to take a long time or require extensive resources should be run inside of a job)

  • Building Python environments, such as with Conda, UV, or Venv.

  • Downloading files (for large data transfers, please request access to the data transfer nodes)

If you need assistance submitting jobs, please contact us.

Account and Account Password Sharing Policy#

No two individuals should share the same HPC account, nor should two individuals share the password of the same account. Each individual is entitled to have their own account hence please request one.

Data Transfer Nodes#

Access to data transfer nodes is available by request and legitimate need. These nodes are meant only for transferring large amounts data to/from HPC resources. They should not be used for computational jobs.

AI Guidance#

We realize that AI can play an important role in research. AI can also impact HPC services, security, data, and compliance with funding agencies, laws, and university policies.

The following are some key points to consider when using AI tools. While we direct these specifically at AI, they also apply to any software, tools, or commands run on SMU HPC systems.

  • Providing an AI with your SMU or HPC credentials or using AI tools to circumvent system logins is a violation of SMU policy.

  • Users are responsible for actions taken by any AI they choose to use. AI tools are often stochastic in nature and may not behave as expected. It is possible that an AI tool may take unexpected actions that adversely impact research (for example, AI tools may corrupt or permanently delete data even if explicitly instructed not to.)

  • Only share data with AI tools that you are allowed to share. Even if you are allowed to share data, we recommend exercising caution, especially if the data is sent to a third party service (these services may use your research to train their models or for other purposes.)

We highly recommend reviewing SMU’s Generative AI Guidance.