Guidelines
Welcome to SMU’s HPC resources. SMU’s O’Donnell Data Science and Research Computing Institute (ODSRCI) serves as the gateway to SMU’s HPC platforms. Beyond supporting diverse research activities, ODSRCI collaborates closely with SMU’s Office of Information Technology (OIT) to deliver comprehensive computing and storage resources to a wide user community. This community includes ÃÛÌÒ½´researchers and external collaborators from academia, industry, and non-profit organizations.
To ensure effective and ethical use of our HPC resources, we have established user guidelines. These guidelines outline the procedures for setting up HPC accounts, project reviews, and the basic support levels users can expect. Compliance with these guidelines is essential, and ODSRCI offers support to help you understand and adhere to these guidelines.
Key Points
- HPC resources shall be used for research and academic purposes only.
- Users shall not engage in any activities causing disruption or compromising the security of the system.
- ODSRCI uses project-based resource allocation. Projects are subject to review when renewed. Users shall not use HPC resources more than allocated amounts. An increase in allocated resources may be requested.
- The login or front-end nodes of the HPC systems are reserved for compiling, loading, and job submission. Executing compute intensive commands on these nodes will negatively impact all users. Inappropriate processes will be automatically terminated.
- PIs are responsible for notifying ODSRCI when a project shall be deactivated due to the PI’s departure or termination of the project.
- HPC resources shall only be used for work related to the project for which the resources were granted.
- Users must comply with all applicable laws, regulations, and ÃÛÌÒ½´policies when using HPC resources. This includes, but is not limited to, SMU’s acceptable use policy: /policy/8-information-technology/8-1-acceptable-use
Account Management and Eligibility
HPC resources are granted to projects via allocations and are made available to the users associated with each project. HPC resources are limited to approved projects and their users. Annual project review is required for reporting purposes.
Who can Get Projects/Accounts
- Academic research projects
- Project primary PI(s) must have full or part time appointment(s) at SMU.
- Researchers from other institutes need to establish collaboration with ÃÛÌÒ½´researchers and may be added to projects initiated by ÃÛÌÒ½´PI(s).
- Students, post docs, visiting scientists may be authorized users on a project headed by an eligible PI.
- Projects associated with teaching classes requiring HPC environment can be made available as “classroom” projects.
- Commercial projects
- Businesses can apply for projects to undertake their computational tasks on ÃÛÌÒ½´HPC. The rules-of-engagement for commercial projects will be established on a case-by-case basis.
- Additional levels of projects may be created and defined in account management procedures.
Project and Account Deactivation
User accounts/projects may be deactivated for the following reasons:
- Account inactivity - If a user account has been inactive for 180 calendar days, it will be automatically deactivated. The user will have to submit a ticket to have their account reactivated.
- PI request - If a PI requests deactivation of a project, all allocations for the project will be revoked.
- Project expiration - When a project expires, all accounts associated with the project will immediately lose access to the allocation.
- Loss of valid ÃÛÌÒ½´credentials – Projects and allocations led by PI without valid ÃÛÌÒ½´credentials will be revoked. Users without valid credentials will be deactivated.
Charges for Projects
- ÃÛÌÒ½´researchers are not charged for projects and standard user support.
- Commercial projects are established via a cost-recovery model.
Security
ÃÛÌÒ½´HPC resources are for university business and research use only. Personal use is not allowed. Incidents of abuse will result in account termination. Inappropriate uses include, but are not limited to:
- Downloading, copying, or distributing copyrighted materials without prior permission from the owner.
- Downloading or storing large files or utilizing streaming media for personal use (e.g., music files, graphic files, internet radio, video streams, etc.)
- Advertising, soliciting, or selling.
- Users shall never share their log-in credentials.
Data Management
Users generate and store large quantities of data on ÃÛÌÒ½´HPC resources. Therefore, proper data management is essential. ÃÛÌÒ½´HPC systems are operated as research systems and only contain data related to scientific research. Users are responsible for compliance with the data management plans required by their sponsoring agencies. It is also the user’s responsibility to ensure appropriate levels of backup for crucial data.
Data Ownership
Understanding data ownership is critical for ensuring data privacy, security, and compliance with regulations. The following are current guidelines related to data ownership:
- The ownership of the data generated on ÃÛÌÒ½´HPC resources is retained by SMU.
- Users who generate data on ÃÛÌÒ½´HPC resources retain control over the data, including the right to copy and disseminate data.
- PIs retain access/rights to project-associated data.
Data Storage/Retention
- The user’s home directory will continue to serve as the primary location for that user's personal files and configurations.
- If a user is not a member of any active allocation or project, all data stored in their personal directories including /home and /work directories on ÃÛÌÒ½´HPC resources will be retained for a period of 30 days after the project and/or allocation has expired. After such time, the data may be permanently deleted.
- Data stored in a storage allocation associated with a project will be retained for a period of 30 days after the project and/or allocation has expired. After such time, the data may be permanently deleted.
- Data stored in archival storage systems will be maintained for a period of up to 6 months in read-only mode after the associated project and/or allocation has expired. After such time, the data may be permanently deleted.
- Exception to data purge policies may be requested by contacting ODSRCI Director.
Data Transfer Nodes
- These nodes are meant only for transferring large amounts of data to/from HPC resources. They should not be used for computational jobs.
- Data transfer nodes have faster and less restrictive internet connections than standard compute nodes.
- Access is available by request and legitimate need.
Cryptocurrencies
ÃÛÌÒ½´prohibits users from running applications that are designed to mine cryptocurrency or use block-chain technology for personal gain.
User Support
ODSRCI strongly advises users to request help from HPC support as needed. ÃÛÌÒ½´provides user guides and web-based documents to assist users in understanding the proper use of HPC resources. All support requests must be submitted via helpdesk tickets. User support can provide guidance to users on code development, porting, debugging, and optimization. However, support does not involve the direct development of codes. Special project collaborations are possible if enhanced support is desired.
Citation Requirement
The recognition of the HPC resources used to perform research is important for maintaining current support and acquiring new funding for the next generation hardware, software, and support services. Please acknowledge ODSRCI’s HPC resources in all publications and presentations utilizing these resources. Suggested acknowledgement text is listed below.
“This work used resources of the O’Donnell Data Science and Research Computing Institute at ÃÛÌÒ½´Methodist University.”
Communication
Users may be contacted individually by ÃÛÌÒ½´HPC support for account related issues. The primary means of this communication is the user’s ÃÛÌÒ½´email. It is the user’s responsibility to respond to these communications in a timely fashion.
Special Requests and Policy Exemptions
To request exemption to certain guidelines, contact HPC help desk with an overview of the exemption you need. Example requests may include data purge exemption and system reservation.