The Texas Advanced Computing Center (TACC) at the University of Texas at Austin moved to the next phase of the planning process for the Executive Class Data Center (LCCF)a process that has many stages of approval and will take about four more years.
If ultimately selected for construction, the LCCF will serve as the leading advanced computing facility for the US academic open science community, helping to solve the most computationally intensive problems. In March 2022, the peer review panel organized by the National Science Foundation (NSF) strongly recommended that the project move to the final design phase.
“The LCCF will be a unique facility,” said Dan Stanzione, executive director of TACC and associate vice president of research at UT Austin. “It will have a broader mission than the Department of Energy and will support a wider range of codes. And it will be more science-focused than the commercial cloud can be, with the hardware, software and—most importantly—the people needed to make discoveries. We look forward to the positive feedback from the panel and look forward to moving on to the next phase of design and planning.”
Once complete, thousands of scholars and students will interact with the LCCF each year. The main actors of the LCCF are US scientists who perform large-scale modeling and simulations; Researchers at major scientific institutions such as LIGO and the Large Hadron Collider; AI and machine learning researchers; and the broader cyber infrastructure community, including staff from other supercomputing centers.
TACC announced during its preliminary design presentation that the LCCF advanced computing system will likely be hosted in a switch commercial data center under construction on the Dell Round Rock campus, 10 miles north of TACC. A new facility at the JJ Pickle Research campus in Austin, in a building adjacent to TACC’s current data center, is also under consideration. TACC leadership estimates the co-housing will save up to $80 million in construction costs and allow the LCCF to support science nine months sooner
A final decision on the main computer architecture for TACC’s next-generation supercomputer, which will be called Horizon, has not yet been made. TACC’s leadership presented their vision of a holistic, collaborative computing ecosystem that supports simulation and modeling as well as machine and deep learning approaches, while offering a variety of services and capabilities to cover the entire data lifecycle – from experimental discovery through analysis to publicly released datasets.
Partnering with HPC leaders
As part of the project, the LCCF will partner with four Distributed Science Centers that will provide unique expertise across the national CI ecosystem and expand support for researchers:
- The partnership with Atlanta University Center (AUCC) Consortium will use five Historically Black Colleges & Universities (HBCU) to provide employees with diverse pathways into computer science and data science.
- The partnership with National Center for Supercomputing Applications (NCSA) will research new processor chips for artificial intelligence (AI).
- The partnership with Pittsburgh Supercomputing Center (PSC) will provide data-intensive computing and data mirrors for published archives, focusing on protected data and FAIR (Findability, Accessibility, Interoperability and Reuse) access.
- The partnership with San Diego Supercomputer Center (SDSC) will provide testbeds focused on supporting machine learning inference in scientific workflows, exabyte-sized instrument data workflows, and methods to democratize access to a broader community.
In preparation for the LCCF, TACC directs the newly introduced ones Feature Scientific application Program. With Support from the NSFthe program is Funding for 21 research teams Working with TACC to improve scientific software, algorithms and libraries.
The selected applications come from a wide variety of fields and use a range of computational approaches, from molecular modeling to deep learning. Together they form a representative sample of the research workload that TACC is expected to support with the LCCF. They will also form the basis for a series of benchmarks against which improvements of the future system over Frontera will be measured.
Reach the missing millions
In addition to providing computing resources and support, the LCCF will build on proven educational and outreach programs that TACC has developed over the past decade. These include SPICE UP, a data science curriculum developed in collaboration with Chaminade University in Hawaii; that Advanced Computing for Social Change Institute, a week-long hackathon for college students emphasizing social justice and data science solutions; and [email protected]Residential summer programs that teach underserved high school students computer science and engineering.
TACC will also expand its internship program, which has successfully brought non-traditional candidates into the HPC field. A new learning center at TACC will serve as a hub for public engagement, training and outreach.
Kelly Gaither, Director of Health Analytics at TACC, said, “The LCCF will not only serve as a research accelerator for the nation, but will also be an agent of change, helping to channel the ‘missing millions’ into the STEM and HPC fields bring .”
Planning for the Leadership-Class Computing Facility is supported by NSF Award #1940979.
About the author
Aaron Dubrow is a science and technology writer with the Communications, Media & Design Group at the Texas Advanced Computing Center.
Originally published on the TACC website.