About PARCC
On this Page
Description
The Penn Advanced Research Computing Center, aka PARCC, is a critical hub for cutting-edge computational research. PARCC will provide resources, expertise, and support to faculty and researchers across diverse disciplines.
PARCC is intended to serve as a centralized computational resource available to all 12 schools at Penn. This will include support for system administration and data science consulting services that cater to our diverse research computation needs. The infrastructure will be hosted at a third-party co-location data center, providing the flexibility to scale up or down as needed while taking advantage of economies of scale and Penn’s purchasing power. The staff will be housed in CCB. Through implementing a sustainable operational model, PARCC aims to ensure future-proofing of the workforce, hardware upgrades, deployment of new technology, and efficient resource allocation. Oversight of this computational resource will be managed by a robust institutional governance model, ensuring fair and equitable access for all researchers, including resource allocation, job scheduling, and data storage.
PARCC will offer general high-performance computing (HPC) and a condo model for faculty who prefer owning their hardware but can utilize co-location with more extensive resources. Hot data storage co-location will increase data transfer speeds and cost efficiency.
In 2023, a diverse faculty working group evaluated Penn’s computational research activity and resources, ultimately producing the summary report titled Positioning Penn at the Frontier of Data and Information-Driven Scholarship. In response to the rapid evolution of data analytics and computational technologies, the group recommended centralizing physical infrastructure to support academic research trajectories. They also proposed the development of a central data science technology resource to provide support, including a dedicated staff pool comprising system administrators, software engineers, and data scientists. As a result of this assessment and recommendation, PARCC is being established as an institutional-scale research computing center.
Looking ahead, Phase II of PARCC will provide additional resources to the computational research community on campus, including data science consulting and additional HPC services and increased capacity based on the need and growth of our computational research community.
Core Principles of PARCC
I. Create a shared, generalist, computational resource for Penn’s 12 schools
II. Provide equitable access to all researchers, including fair resource allocation, job scheduling, and data storage
III. Design infrastructure that supports diverse needs balancing computational power (CPUs & GPUs), storage capacity (hot & cold), and network bandwidth
IV. Hire a central workforce to support system administration needs and provide data science consulting services to researchers
V. Advance a sustainable operational model that future proofs workforce, hardware upgrades, new technology deployment, and efficient resource allocation
VI. Develop a governance model that provides oversight to the core investors in this new computational resource
Phase I Timeline
- May – June: Requirements
- Identify research requirements and computational tasks
- Conduct stakeholder reviews
- June – July: Planning and Design
- Network requirements and data center specifications
- Software and tools for various research disciplines
- August – October: Budget and Funding
- Develop financial models
- August – October: Procurement
- Select vendors and negotiate requirements
- November – July
- Install and Configuration
- Install data center, hardware, software
- Security configurations
- Testing and Optimization
- Initial testing and benchmarking
- Optimize configurations
- Pilot tests with a select group of researchers
- Training and Support
- Conduct training for researchers
- IT support structure and helpdesk
- Security and Compliance
- Implement data security measures
- Ensure compliance with institutional and regulatory standards
- Monitoring and Evaluation
- Performance monitoring
- User feedback through surveys and forums
- Plan for updates and enahancements
- Install and Configuration
Related FAQs
How do I know my data is secure?
Data security is very important to us at PARCC. We are working with our vendors and ISC Compliance office to ensure that we have robust data security. Details on security and compliance specifications will follow.
How much will it cost to use the Cluster?
We are working to find a model that will enable the growth of the center while being financially responsible for our researchers to utilize. We will provide guidance for common cost comparisons around the university. Sign up for our mailing list to stay updated!
If I have a question about PARCC, whom do I contact?
Please contact us at parcc-info@parcc.upenn.edu
Until PARCC is available, can I purchase equipment to use now and merge it when it’s ready?
Yes! While we don’t know exactly what this will look like yet, we are planning on supporting these sort of options. Please reach out to parcc-info@parcc.upenn.edu and your School IT department to coordinate with the latest information.
What kind of training is available?
Initially, we will have a knowledge base to help users get started. As we build out capabilities, we hope to introduce more thorough training options. Please email parcc-info@parcc.upenn.edu with any specific questions.
What support services will be available?
We aim to offer support surrounding your research journey using the PARCC system. As we build out capabilities, we will have more explicit support offerings. Please email parcc-info@parcc.upenn.edu with any specific questions.
When will PARCC be available?
PARCC aims to be operational in Spring 2025.
Where is PARCC located?
The exact location of PARCC is yet to be determined, but will be in the mid atlantic region to provide a highly responsive experience from campus
Who will be able to use the PARCC HPC cluster?
PARCC will have a phased approach with a pilot during 2025 and opening up for more general availability to Penn researchers after the pilot is completed.