The Anvil login and GPU (Nvidia A100 GPUs) nodes, along with AnvilGPT, Anvil Notebook, and storage will stay online.

How Does This Affect You? Anvil CPU and AI nodes, along with Anvil Composable, will be unavailable during all outage windows. Jobs on CPU and AI nodes will not run, and new jobs to CPU and AI nodes will not start while they are offline. Jobs on CPU and AI nodes that cannot be completed before an outage begins will not start. They will remain in the queue until these nodes returns to service.

Any Questions? If you have questions about these outages or need assistance, please submit a ticket through the ACCESS Help Desk.

Neutrinos are fundamental subatomic particles that possess no charge and nearly no mass. Due to these unique characteristics, neutrinos do not interact with electromagnetic forces or the strong nuclear force, and rarely with gravity or the weak force. In fact, neutrinos typically pass directly through matter, including Earth, with no discernible effect. This makes studying the particles extremely difficult. However, the elusivity of neutrinos is exactly what makes them worthwhile for study—because they interact with so little in the universe, they carry with them unaltered information about the cataclysmic astrophysical events (black hole formation, Big Bang, etc.) that created them. As Halzen himself put it in a 2014 Astronomische Nachrichten publication:

“High-energy neutrinos have a distinct potential to probe the extreme Universe. Neutrinos reach us from the edge of the Universe without absorption and with no deflection by magnetic fields. They can escape unscathed from the inner neighborhood of black holes and from the accelerators where cosmic rays are born.”

As noted, most neutrinos pass directly through the planet. Some, however, do occasionally collide with other subatomic particles, causing an interaction that leaves a trace. To view these interactions, and therefore to observe and study neutrinos, one must essentially set up sensors, then sit and wait. Scientists have known this since the early 1970s and determined that a particle detector at least one cubic kilometer in size would be required for the task. In the late ‘80s, Halzen realized the ideal location for such a detector would be inside ice, where sensor noise would be much lower than anywhere else. Thus the idea of the IceCube Neutrino Observatory was born.

The NSF approved initial funding for the IceCube project in 2002, and construction of the facility at the South Pole was completed in 2010. Soon after, results began rolling in, and with them, massive amounts of data. Currently, the IceCube observatory produces over one terabyte of data daily. This data then needs to be analyzed in order to derive any meaningful insights, which requires a substantial amount of HPC resources. To deliver both the data and the HPC power to researchers, the IceCube project relies on the NSF’s advanced computing portfolio, including the supercomputing resources coordinated through the ACCESS program.

“Anvil and ACCESS have been strong partners with IceCube for many years, supporting a diverse range of scientific analyses,” says David Schultz, Research Cyberinfrastruture Manager at the Wisconsin IceCube Particle Astrophysics Center. “ACCESS GPU resources are critical for accurately modeling the glacial ice at the South Pole, a necessity for understanding IceCube data.”

Purdue’s Anvil supercomputer has contributed significantly to the IceCube project via its ACCESS allocation. To date, IceCube has used nearly 160 million SUs (service units) on the system, the second-highest ACCESS machine usage for the project. While this is welcome news for the Anvil team, it was also not a surprise—IceCube is precisely the type of large-scale compute project that Anvil was designed for. The cluster boasts over 1000 compute nodes as well as 37 dedicated GPU nodes, 21 of which contain 4 NVIDIA H100s each. No matter the type of job needed by IceCube researchers, Anvil’s system can provide.

Purdue also supports the IceCube project by contributing HPC resources aside from Anvil to the Open Science Pool (OSPool). OSPool is a distributed high-throughput computing platform that connects resources from universities and national labs and organizes them as a single, virtual cluster. The IceCube team has run nearly 120,000 jobs thus far on Purdue systems via OSPool.

“Congratulations to Dr. Halzen for being awarded the Nobel Prize,” says Preston Smith, Executive Director of the Rosen Center for Advanced Computing (RCAC) at Purdue. “The IceCube team has achieved the extraordinary, and we’re proud to have our HPC resources contributing."

More information about the 2026 Nobel Prize in Physics can be found here: Nobel Prize Press Release. For more information on Dr. Halzen’s work and the IceCube project, please visit: IceCube Neutrino Observatory

To learn more about high-performance computing and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

Five students from across the nation gathered at Purdue’s campus in West Lafayette, Indiana, for the 2026 CI-XP Summer Internship Program. These students joined the program to learn about HPC and work on projects involving Purdue’s supercomputing resources, including the NSF-funded Anvil supercomputer. Over the course of the internship, the students gained the knowledge and skills required to build advanced research computing systems and scientific applications to run on them.

The CI-XP Summer Internship Program is a paid summer internship open to undergraduate students in the United States, regardless of their background. As in previous years, a massive influx of applications caused the application window to close nearly a month before its official deadline, with total applicants surging to nearly 400 students. The CI-XP mentors—five RCAC staff members who led the projects and guided the students through the internship—narrowed the list down to just five students. The five participants of the 2026 CI-XP Summer Internship Program were:

• Branden Turner, Computer Science major, Texas A&M University
• Eric Jin, Computer Science major, Purdue University
• Howie Lin, Computer Science and Economics double major, Boston College
• Angello Bravo, Computer Science major, New Jersey Institute of Technology
• Shelby Stall, Computer Science and Business Analytics double major, DePauw University

The CI-XP Summer Internship Program consisted of three separate projects. These projects were chosen with real-world applicability in mind—the students would not only gain experience with HPC and learn new skill sets, but would simultaneously increase RCAC’s system capabilities. Each student also had a personal project mentor to help them achieve their goals.

Project 1:

The first CI-XP project for 2026 focused on building the GTAP Insight Engine, a next-generation AI-powered tool designed to make the outputs of the GTAP (Global Trade Analysis Project) model more accessible to researchers and policymakers. Two students, Howie Lin and Eric Jin, teamed up to take on this project under the supervision of their mentors, Haniye Kashgarani and Mihir Ahlawat, as well as RCAC Executive Team member Arman Pazouki. Erwin Corong, Deputy Director and Principal Research Economist for GTAP, also assisted in mentoring the students and helping with their work. In this project, Lin and Jin built a tool that utilizes AI agents to interpret natural language queries, extract relevant data from GTAP model outputs, and perform basic data analysis. The GTAP model is used extensively to analyze issues related to international trade, energy policies, agriculture, and the environment, but its 800+ model variables make it daunting for novice researchers. Lin and Jin tackled this problem by designing an AI tool that allows users to interact with GTAP’s complex model outputs using plain language and presents all results in a user-friendly visual format. Their AI tool was developed using a tech stack that included AnvilGPT (Purdue’s on-premises LLM service), LangChain​, PostgreSQL​, ChromaDB​, and Chainlit​. They incorporated a two-step pipeline consisting of a “planner” agent and a “worker” agent, included Retrieval-Augmented Generation capabilities, and added AI-agent-powered data collection and analysis features. By the end of the summer, Jin and Lin delivered a well-functioning GTAP Insight Engine that lowers the barrier to entry for exploring complex economic model results. The tool is currently deployed through the Anvil Composable Subsystem.

“I came in with some coding experience, but not enough to fully make the agent without more learning,” says Lin. “I think that is definitely a huge benefit of being part of this program. For somebody who is just interested in coding and has a bit of background in it, it gives you a lot of time to learn more about the code.”

Project 2:

The second CI-XP project worked on developing Elastic Cloud Scheduling for Edge Kubernetes. Angello Bravo and Shelby Stall worked on this project under the supervision of their mentors, LJ Lumas and Frank Bakhit, and RCAC Executive Team member Erik Gough. At the onset of the summer, RCAC’s HPC resource scheduler could only convert entire GPU nodes into Kubernetes workloads, meaning smaller cloud computing jobs may only utilize a single GPU at a time, but would still have 4 allocated, leaving 3 GPUs idle. This is an inefficient use of the highly coveted GPUs available to Purdue researchers. To solve this, Bravo and Stall developed a fine-grained scheduling system that enables single-GPU allocation across heterogeneous compute environments and multiple Kubernetes clusters. Their development pipeline included tools such as InterLink, Virtual Kubelet, Kubewarden, and Helm. Thanks to their work, Purdue researchers can now use powerful GPUs more efficiently while reducing idle GPU time, thereby increasing accessibility for AI, machine learning, and data-intensive research.

“I really enjoyed working in a research environment,” says Stall. “Everyone was super supportive and helpful and really open to answering any questions that we had, and really wanted to help us improve in any way that they could.”

Project 3:

For the third CI-XP project, Branden Turner worked with his mentor, Guangzhen Jin, and RCAC Executive Team member Preston Smith, to make Open Science Data Federation (OSDF) datasets discoverable on Purdue’s Anvil supercomputer through Pelican, a software platform that federates and delivers data repositories to researchers throughout the nation. Turner’s first step was to run three benchmark tests to determine whether Pelican would be useful for data streaming to researchers on Anvil. During these benchmarks, Turner found that Pelican would indeed be useful, but that there were currently two barriers that could be removed: discoverability and streamability. The discoverability barrier fell to the fact that OSDF does not have a central dataset catalog, so researchers needed to know specifics of what they were looking for to find it. For streamability, Turner found that certain compressed data packages posed issues in streaming reliability. To solve both of these problems, Turner developed PELICANVIL. PELICANVIL is built on a tech stack that includes FastAPI, PelicanFS, and SQLite. It provides manually indexed datasets, a dataset browser, a quick access panel, and an admin panel, all while improving streamability as well as allowing for downloads. Researchers can access PELICANVIL with a web browser through the Anvil Open OnDemand dashboard, which further lowers the barrier to entry for new users. Thanks to Turner’s work during the CI-XP program, PELICANVIL significantly improves performance for data-intensive workloads, including large-scale simulations, AI/ML training, and high-throughput analytics, and enables a fast and reliable research pipeline across multiple scientific disciplines. The paper from this project, titled, "Pelicanvil: A Web Platform for Discovering, Browsing, and Accessing Federated Datasets on HPC," has been accepted at SC26. Turner will be presenting the work to the broader HPC community at the conference held in Chicago on November 15.

“This project showed me how important software development is for research and especially research that uses high-performance computing,” says Turner. “Without this program, I wouldn't even know about HPC or RSE [Research Software Engineering].”

A comprehensive educational experience

While the CI-XP students worked hard all summer, this program was more than a temporary job—it was a completely immersive learning experience. The students had access to their mentors and to the RCAC staff working on-site. They were also able to take tours of the campus and data center, attend presentations hosted by both RCAC and the SURF (Summer Undergraduate Research Fellowship) program, participate in technical workshops, and even attend and present at the Practice and Experience in Advanced Research Computing (PEARC) 2026 Conference! The five CI-XP students also gave midpoint presentations on their projects to staff from Oak Ridge National Laboratory (ORNL), the Institute for Geospatial Understanding through an Integrative Discovery Environment (I-GUIDE), the Morgridge Institute for Research, and Pacific Northwest National Laboratory (PNNL).

“PEARC was amazing,” says Bravo. “The work there was super cool. I met so many professionals and students and was able to attend many sessions and learn more about not only supercomputing, but also systems and apps.”

Lin adds, “I learned a lot at PEARC. After spending the weekend there participating in student program events, you get to explore all the sessions, read some of the best papers, and hear from some great speakers (as part of the full conference). You get to talk with companies and possibly get yourself an internship.”

Beyond technical workshops and presentation experience, the REU students took part in workshops focused on developing intangible skills. One such workshop, led by Syd Moore, an Academic Advisor & Gallup-Certified Strengths Coach, alongside Manabu Taketani, a Leadership Development Specialist at Purdue LEAD, guided the group through the myStrengths talent assessment. Another was led by Matt Jones, Certified Master Facilitator in LEGO® SERIOUS PLAY® methods (LSP). This session introduced the students to LSP, a facilitated thinking, communication, and problem-solving technique for use with organizations, teams, and individuals. Both workshops took place at the beginning of the summer, followed by a midpoint meeting to build on the lessons learned and further develop the students for their future careers.

“Some events that stood out to me were the first day when we did some LEGO series, play, team-building activities, and strengths workshops to better understand how we would work together as a team throughout the summer,” says Stall. “This was a great introduction to the program, and it was really helpful to get to know each other before jumping into these projects.”

Stall continues, “It was also really impactful to revisit what we had been doing throughout the summer and what we did on the first day after the midpoint presentations to kind of build on what we worked on to be able to continue and improve our team building and collaboration throughout the summer.

Giving these students the opportunity to gain as much knowledge and experience as possible is a vital component of the CI-XP program. In this way, RCAC helps ensure that each summer program participant develops into a capable and competent cyberinfrastructure professional.

Mission accomplished

On the final day of the CI-XP program, the students presented their work to the RCAC team. As they demonstrated the results of their projects, each student discussed their accomplishments, obstacles, failures, and what they learned throughout the summer. The students were then asked questions and given feedback on their presentations. To the RCAC team, it was wonderful to hear how the summer program might help steer the future careers of these students, many of whom expressed a desire to continue within the field of HPC.

“This CI-XP program has been very insightful for someone like me who has looked more towards industry-based careers, not towards research,” says Turner. “It really opened my eyes to how high-performance computing is really useful to research.”

Overall, these five students made fantastic progress: they completed their projects, learned technical and interpersonal skills they will need in the workforce, and gained an in-depth understanding of the HPC world. To sum up the summer, Bravo shared this quote:

“I would 100% recommend this CI-XP program to other students because having this type of experience as an undergrad is very complicated. It's a very cool opportunity.”

To learn more about the 2026 summer CI-XP program, please visit our Research Experience for Undergraduates webpage. Registration for 2027 will open once the new projects are announced in early November, so be sure to bookmark the page and keep an eye out on RCAC’s social media channels. The application deadline is February 19, 2027, but may close earlier based on the volume of submissions. Interviews for positions will begin in January of 2027.

For more information regarding HPC and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

The Big Data Training for Cancer Research (BigCARE) workshop is a program funded by the National Cancer Institute (NCI). It was founded in 2020 by Min Zhang, MD, PhD, a Professor of Epidemiology and Biostatistics at Wen Public Health and the Biostatistics Shared Resources Director for the UCI Chao Family Comprehensive Cancer Center. Her founding collaborators include: Dr. Sean Davis, MD, PhD, Associate Director of Informatics and Data Science, Professor of Medicine, from the University of Colorado Anschutz School of Medicine, and Dr. Dabao Zhang, PhD, Professor of Epidemiology and Biostatistics at Wen Public Health. Recognizing the need for specialized HPC and Big Data training for cancer researchers, the team designed BigCARE to provide for that need. This year’s workshop focused on teaching 1) the basics of R programming; 2) data analysis workflows for multiple cancer data types, including targeted and untargeted metabolomics data, bulk and single-cell RNA-Seq data, ChIP-seq and ATAC-seq epigenomics data, and microbiome data; and 3) the basic concepts for conducting gene association studies, genotype-instrumented causal inference, and causal gene regulatory networks construction.

“Anvil has been extremely helpful during the previous BigCARE workshops,” says Zhang, “especially for our participants with limited computing skills. Anvil provides the essential infrastructure and computing support needed to navigate between command line and R packages for large-scale data. This year, Anvil made the implementation much smoother when we added some AI and machine learning tools for multi-omics data analysis. The Anvil platform, along with Jupyter Notebook, offered an all-in-one solution that helped participants easily and quickly switch from concept to interactive analysis of big data without obstacles.”

Anvil’s role in the BigCARE workshop was to provide HPC resources through Open OnDemand and Jupyter Notebooks, which limited the need for in-depth knowledge of command-line interfaces or HPC server environments. The course material was developed as Jupyter notebooks, which researchers could access directly via the web thanks to Open OnDemand. All of this equated to a low barrier of entry for the workshop participants.

For BigCARE 2026, staff members from the Rosen Center for Advanced Computing (RCAC) pre-installed all the bioinformatics tools and datasets that participants and instructors used for the workshop, resulting in a smoother workshop experience for all. Eric Adams, the Lead Research Operations Administrator for Education, and Ryan DeRue, a Senior Computational Scientist, also attended the event at UCI to serve as instructors for the HPC/Linux portion of the workshop and provided on-site support.

“This is my third year serving as an instructor and providing on-site support for the BigCARE workshop, and each year I've attended, I am always struck by how impressive the attendees of the workshop are,” says DeRue. “They are comprised of faculty, med students, postdocs, and staff, and each of them is extremely knowledgeable in their focus area within cancer research. So, it's especially fulfilling to me to play a part in connecting their ideas to a system like Anvil where those ideas can culminate. Many of the attendees go on to become computational researchers in their own right, and so I've always viewed the BigCARE workshop as mutually beneficial for the attendees as well as for us. In some sense, we get to cultivate the computational intuition of the researchers that will one day be applying for their own allocation on our systems.”

As in years prior, the BigCARE 2026 workshop was a great success. Dr. Zhang and the attendees were excited by what they accomplished during the two-week intensive, and pleased with how helpful the RCAC support team was. In a post-course survey, 17 participants said they were likely or very likely to apply for their own Anvil allocation in the future, with 27 planning to continue using the system for the remainder of their one-year allocation, which was included as part of the workshop. Dr. Zhang also indicated that she intends to continue using Anvil to support training activities for the foreseeable future.

More information about the BigCARE 2026 Summer Workshop can be found on UCI’s “Big Data Training for Cancer Research” webpage. Information about the Anvil supercomputer can be found on Purdue’s Anvil Website.

For more information regarding HPC and how it can help you, please visit our “Why HPC?” page. Anvil is funded under NSF award No. 2005632. Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process.

Written by: Jonathan Poole, poole43@purdue.edu

The scheduled maintenance on Anvil has been successfully completed as of 3:37 pm EDT. All services are now available. The queues have been enabled and jobs have resumed scheduling. If you encounter any issues following this maintenance, please contact Anvil team through ACCESS Help Desk.

**Updates (September 18, 2026): **

This is a reminder that Anvil is scheduled for a power outage and maintenance starting tomorrow, Saturday, September 19, at 12:00 a.m. through Monday, September 21, at 8:00 p.m. EDT.

Updates (September 10, 2026):

This is a reminder that Anvil is scheduled for a power outage and maintenance on Saturday, September 19, at 12:00 a.m. through Monday, September 21, at 8:00 p.m. EDT.

Updates (August 13, 2026): The power outage originally scheduled for Saturday, August 15, at 12:00 a.m. through Monday, August 17, at 5:00 p.m. EDT has been cancelled, therefore Anvil remains available during this time.

Anvil is still scheduled for a power outage and maintenance on Saturday, September 19, at 12:00 a.m. through Monday, September 21, at 8:00 p.m. EDT.

Original post (August 7, 2026): The Anvil clusters and services will be unavailable during two scheduled maintenance events in August and September to support datacenter improvement projects that require power outages.

The first power outage is scheduled for Saturday, August 15, at 12:00 a.m. through Monday, August 17, at 5:00 p.m. EDT.

The second power outage and Anvil maintenance are scheduled for Saturday, September 19, at 12:00 a.m. through Monday, September 21, at 8:00 p.m. EDT.

The Anvil quarterly maintenance previously announced for August 25 has been cancelled, and Anvil willl remain available on August 25.

How Does This Affect You?

Anvil CPU, GPU, and AI clusters, along with AnvilGPT, Anvil Notebook, Anvil Composable, and storage, will be unavailable during all outage windows. Jobs will not run, and new jobs will not start while Anvil is offline. Jobs that cannot be completed before an outage begins will not start. They will remain in the queue until Anvil returns to service.

Any Questions?

If you have questions about these outages or need assistance, please submit a ticket through the ACCESS Help Desk.

As the renovation and capacity improvement project in the MATH datacenter moves into its final stages, a planned maintenance outage is required to continue bringing the newly installed power and cooling infrastructure online. This work is part of a broader renovation that will result in a 32% increase in floor space, a 60% increase in usable power, and a two-fold increase in cooling capacity — all crucial facility ingredients needed to continue providing world-class high-performance computing resources to Purdue faculty and the nation.

**Outage Window: **Saturday, September 19, 2026, 12:00 AM – Sunday, September 20, 2026, 6:00 PM

What will remain online

The following resources are not affected by this power work and will remain available throughout the maintenance window:

• Data Depot and Data Depot Object
• ScienceDMZ
• Scholar
• Fortress
• HubZero
• Hammer
• Weber

What will go offline

The following resources, including their associated Data Depot storage and Data Depot Object allocations, will be unavailable during the outage window and will return to service gradually once the power work is complete:

• GenAI Studio
• Geddes
• Anvil
• Gautschi
• Gilbreth
• Bell
• Negishi
• Rowdy
• Rossmann

How does this affect you?

Jobs on the affected clusters will not run, and new jobs will not start during this window. Jobs that are expected to not complete before the maintenance begins will not start until after the maintenance window ends. Resources listed above as remaining online will continue to operate normally throughout.

Questions

If you have questions about this outage or need help from our support staff, please contact rcac-help@purdue.edu.

Best, RCAC

PEARC is a renowned conference dedicated to advancing high-performance computing (HPC) in research worldwide. For PEARC26, a record-breaking 1000+ attendees gathered in downtown Minneapolis — a fitting benchmark to celebrate the conference’s tenth anniversary. Throughout the week, attendees participated in plenary sessions, tutorials, workshops, poster sessions, panels, and research presentations. The conference even hosted a student program, which was extended to a full seven days this year thanks to its overwhelming success in years prior.

Leaders in Research Computing

RCAC staff members didn’t just attend the conference, they helped lead the charge for advancing HPC with multiple papers, talks, and workshops. A full list of presenters is as follows:

• Suzanna Gardner: Resilience Is a Team Sport: Strategies for Cultivating Resilient Roots and Building Sustainable Research Computing Teams
• Haniye Kashgarani: AnvilOps: Increasing Kubernetes Accessibility via an Open-Source Platform-as-a-Service
• Jaewoo Shin: Toward a Trustworthy and Accessible Scientific Data Workflow Platform with StreamCI
• Eric Adams: ACCESS-ing the Future: Survival Guide for 2026 and Beyond
• Carol Song: Science Gateways in the Agentic Era
• Geoffrey Lentner: Addressing the Challenges Posed by Anaconda's New Licensing Model
• Guangzhen Jin: Modernizing HPC Documentation: Design, Automation, and Lessons Learned at Purdue RCAC
• Preston Smith: GPU Workload Analysis of Big 10 Campus, Regional, and National HPC Systems
• I Luk Kim: SciOn: An AI-Friendly Framework for Reliable Scientific Application Development
• Suzanna Gardner and Laura Theademan: Building Inclusive Cyberinfrastructure Pathways: Scaling Outreach, Workforce Development, and Community Engagement Through Anvil
• Erik Gough: Dynamic Node Manager — An Autoscaler for Kubernetes Workloads on HPC Resources
• Jungha Woo: A Generalized Methodology for Evaluating AI-IntegratedResearch Software: Lessons from a Smart Search Implementation for Cyber-Training
• George Takahashi: CollabXR: immersive spatial communication of 3D materials (Best Visualization Winner)*
• Geoffrey Lentner: Hello Computer: HPC in the Agentic Era (Best Paper Nominee)
• Jaewoo Shin: HearingHub: A Collaborative Web Platform for Multi-Modal Hearing Research Data Management and Visualization (Best Paper Nominee)
• I Luk Kim: HydroFlame: A Cloud-Native Geospatial Platform for Assessing Post-Fire Hydrological Impacts on HPC Infrastructure

*Congratulations to RCAC’s George Takahashi and Jake White, who won the PEARC26 Best Visualization award for their work, CollabXR: immersive spatial communication of 3D materials. The Best Visualization Award is given to authors who best showcase research data, results, or concepts in a captivating and dynamic format. This award recognizes effective communication and polished presentation, making an impact for both domain experts and the broader community. Great job, George and Jake!

It should be noted that while the individual(s) who spoke at PEARC26 are listed above, each presentation/paper was a collaborative effort involving numerous RCAC staff members. For a full list of Purdue-affiliated presentations, including a list of the co-authors, please visit: PEARC26 Schedule, and search “Purdue”

PEARC26 Student Success

RCAC’s core mission is to provide researchers access to leading-edge computational and data storage systems and expertise in a broad range of HPC activities, as well as to pursue technological innovations that support research computing. As a major center operating at Purdue University, RCAC’s secondary goal is to enhance student and educational outcomes in a variety of ways, including through hands-on workforce development. The center’s CI-XP (Cyber Infrastructure - eXPerience) Program is designed as the workforce development pipeline for HPC at Purdue. CI-XP students gain real workplace experience by taking on tasks with real-world impact — no “toy” projects. RCAC also supports the CI-XP students by sponsoring their attendance at major HPC and research computing conferences. For PEARC26, RCAC sent ten CI-XP students to learn, network, and participate in the PEARC student program. These students included: Eric Jin, Angello Bravo, Branden Turner, Howie Lin, Shelby Stall, Madiha Fathima, Shubham Jha, Amari Gardner, Banon Monroe, and Nividetha Nerella. Travel support was provided by the PEARC26 Student Program for a majority the students. Of the ten who attended, six gave presentations on their RCAC projects, gaining valuable experience in showcasing their work in a professional setting.

“I had such a great experience at PEARC26 and was able to learn from the professionals and students around me,” says Shelby Stall, a participant in RCAC’s 2026 CI-XP Summer Internship Program. “I attended multiple workshops, networking events, and participated in the student program, where I met other like-minded and hardworking students. The HPC community was incredibly welcoming and supportive, which made the experience even more meaningful. I especially appreciated the opportunity to work with my mentor, Manasa Vemuri, who provided great advice and guidance throughout the conference. Overall, PEARC26 was an amazing opportunity to learn, connect with others, and grow both personally and professionally!”

For more information about student opportunities and how the conference supports future workforce development, please visit: PEARC26 Student Program

Enabling Excellence at PEARC26

RCAC also had a major impact at PEARC26 by supporting the conference overall. Anvil, one of Purdue University’s most powerful supercomputers, is an NSF-funded HPC resource designed to provide advanced computing capabilities to researchers nationwide. The Anvil supercomputer was used to host multiple tutorials and workshops at the conference, including: the HPC for Health Sciences workshop, the AI/ML Workflows Across National Cyberinfrastructure tutorial, the AI Workflows on Campus and National Cyberinfrastructure tutorial, and a PEARC26 Student Program workshop.

Aside from computing resources, RCAC supported PEARC26 through active involvement in conference leadership positions. A list of committee members from RCAC includes:

• Katy Gunderson, Executive Committee, Graphic Designer
• Geoffrey Lentner, Technical Committee, Applications: Scientific Impact, Deployments, and Workflows Co-Chair
• Amanda Warren-Glowe, Student Program Committee
• Amanda Hassenplug, Student Program Committee
• Suzanna Gardner, Student Program Committee and Committee for Accessibility and Belonging
• Taylor Graham, Committee for Accessibility and Belonging

Overall, PEARC26 was a great success for RCAC, underscoring Purdue's role as a national leader in research computing. To learn more about the conference, please visit: PEARC26

RCAC operates the centrally-maintained research computing resources at Purdue University, providing access to leading-edge computational and data storage systems as well as expertise and support to Purdue faculty, staff, and student researchers. To learn more about HPC and how RCAC can help you, please visit: https://www.rcac.purdue.edu/ or reach out to rcac-help@purdue.edu to request consultation.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

Created under the NSF’s Office of Advanced Cyberinfrastructure (OAC), the IDSS program aims to provide support for foundational, national-scale data cyberinfrastructure that facilitates AI- and data-driven science and engineering research and education. IDSS is a broad-spectrum initiative focusing on transdisciplinary and multidisciplinary projects. Projects are divided into three categories: 1) development, deployment, and operation of new systems; 2) enhancement of smaller systems to national scale; 3) planning grants for future potential development. The three projects that RCAC is involved with are in categories 1 and 2:

• Category I: Fabric for AI-Driven Science (FabAID): A $24.5M project led by the Morgridge Institute for Research, the FabAID project will support the life cycle of scientific datasets by integrating data repositories with national-scale computing resources. Researchers will be able to utilize FabAID to seamlessly manage the data-intensive workflows required to conduct AI-driven research. Purdue will integrate archival storage capability into FabAID via the Fortress archive. More information about the project can be found here: Fabric for AI-Driven Science

• Category II: Transitioning the National Science Data Fabric (NSDF) Pilot into a National Operational Cyberinfrastructure and Service for Democratized, AI-Driven Scientific Discovery: This project is led by the University of Tennessee Knoxville (UTK) and will scale the successful NSDF Pilot program into a full production-grade national cyberinfrastructure and service. NSDF solves the problem of data isolation. It connects researchers at institutions of all sizes to scientific data wherever it resides—whether it be at national laboratories, experimental facilities, cloud platforms, or campus clusters—while eliminating data movement issues. Once online, researchers at institutions with limited computing resources will still be able to contribute to scientific progress in domains that rely on large volumes of data. Purdue will lead the integration of Entry Points and coordinate the national operation of NSDF. Purdue will also dedicate Anvil object storage and computational resources, as well as host an NSDF Entry Point infrastructure to support the execution of NSDF data-driven workflows on these resources. The NSDF project was awarded $9M, and more information can be found here: National Science Data Fabric

• Category II: iDLab - A Federated Interactive Discovery Lab for Data-Driven Research and Education: The final project that RCAC is partnered on from this round of IDSS awards is the Interactive Discovery Laboratory (iDLab), led by the University of California, Los Angeles. The iDLab project will transition the OneSciencePlace platform from a pilot to a national-scale, data-centric production service that connects five NSF-supported high-performance computing (HPC) resources and two public cloud platforms. Researchers will have access to local data resources, interactive applications, and a shared data partition, all through an easy-to-use web-based platform. Purdue will integrate its Anvil cluster into the iDLab ecosystem, providing new pathways to HPC cluster access and accelerating interactive data-driven scientific discovery. The iDLab project was awarded $9M in funding. More information can be found here: Interactive Discovery Laboratory

“Professor Michela Taufer at UTK and I first began discussing the IDSS Category II idea in September 2025, and it’s exciting to see that idea reach this milestone,” says Carol Song, Chief Scientist at RCAC and co-Principal Investigator on one of the IDSS projects. “This NSDF IDSS project is a true collaboration, and we’re proud to be part of it. Purdue’s RCAC brings the capacity, operational strength, and technical innovation needed to support researchers and students at scale, and we’re especially excited to contribute to this new NSF data program from the beginning. Seeing RCAC’s role across multiple awards in this round highlights the depth of expertise we have here and the momentum of our growing partnerships with other U.S. institutions.”

A major component of the IDSS program is integration; projects are expected to leverage and interconnect with existing national-scale cyberinfrastructure systems and services, such as those provided through the NSF’s ACCESS and NAIRR programs. Purdue’s Anvil supercomputer serves researchers nationwide through both of these programs. Thanks to Anvil, Purdue was able to capitalize on these new funding opportunities.

“Being a partner on three of the initial round of IDSS awards highlights the importance of having national resources at Purdue and the connective role that it serves,” says Preston Smith, Executive Director of RCAC. “Advanced computing and the expertise delivering it is the bridge between researchers and the broader national ecosystem.”

To learn more about the IDSS program, please visit: Integrated Data Systems and Services. Project solicitations are currently open, with a proposal deadline of July 27th, 2027.

The Rosen Center for Advanced Computing (RCAC) operates the centrally-maintained research computing resources at Purdue University, providing access to leading-edge computational and data storage systems as well as expertise and support to Purdue faculty, staff, and student researchers. If you have an idea or project proposal that you’d like to discuss with RCAC, please reach out to: rcac-help@purdue.edu

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

Mission to Mars

The first camp of the season was the inaugural Mission to Mars Summer Day camps. Taking place on June 15, 16, and 17, the camp was open to incoming 7th-9th grade students, with each day being a singular day camp. Students arrived early on their selected date and were led through a full-day, hands-on experience designed to foster critical thinking and teamwork while teaching them about high-performance computing and STEM principles.

During Mission to Mars, campers stepped into the role of Mission Control during a simulated Mars mission in crisis. Through three interconnected, story-driven STEM challenges, the students investigated real problems, designed creative solutions, and collaborated with peers — all in a race to save the mission. Some of the challenges included a cable-driven motion simulation, a teamwork-focused blind navigation exercise, and developing a functional water-filtration system. The students were also able to experience the Envision Center’s CollabXR platform, a powerful virtual/augmented reality tool in which participants can observe 3-dimensional objects and data in a shared learning space.

“It was an incredible opportunity to help lead Mission to Mars through its pilot year and watch as we had the chance to inspire the first student campers in STEM,” says Banon Monroe, a CI-XP student who helped run the camps. “Seeing the students work together to tackle challenges was exciting, and I can’t wait to help put on the camp again next year!”

Code Explorers

The next camp for the summer was “Code Explorers: Coding and AI.” Taking place from June 21-26, this camp was an immersive, week-long STEM program where the students engaged in coding, data analytics, and environmental science. RCAC hosted this alongside Purdue's College of Science as part of the Summer College for High School Students program. Campers engaged in hands-on projects, such as using microcontrollers and sensors to collect real-world data on factors like temperature, soil moisture, and air quality. They then learned about Python and block-based programming, and how to use coding to analyze and visualize data, giving them the power to transform their observations into compelling stories that highlight environmental challenges or successes in their own community. The students also explored the fundamentals of artificial intelligence (AI) and machine learning.

“The Coding Explorers Camp was an awesome summer experience!” says Nicholas Brasovan, a curriculum developer in the Department of Computer Science at Purdue, and one of the instructors for the camp. “Students explored Python fundamentals for data science, basic agentic programming, computer architecture, cloud computing, and ethics of technology. We had a series of hands-on labs hosted at the Lawson Computer Science building. We also partnered with RCAC and the Envision Center to deliver a gamified workshop on data science. This is our third iteration of the camp, and it just keeps getting better.”

CyberSafe Heroes

The final camp of the summer, another summer college program that ran from July 12-17, was “CyberSafe Heroes: Empowering High School Students for a Secure Digital Future.” This camp focused on cybersecurity, introducing students to key concepts in ethical hacking, encryption, and online safety. Through hands-on labs and engaging activities, including cybersecurity escape rooms, ethical hacking simulations, and encryption challenges, the students explored fundamental cybersecurity topics and were equipped with essential digital skills. The campers also explored AI’s role in modern cybersecurity, learning how the tool is leveraged for both noble and nefarious purposes. Overall, the course provided an engaging and educational experience that helped inspire the cybersecurity professionals of tomorrow while empowering students to be secure in the digital world of today.

“CyberSafe Heroes was a great opportunity to introduce students to the exciting world of cybersecurity through hands-on learning,” says George Bailey, the director of cyberTAP at Purdue and one of the instructors for the CyberSafe camp. “From ethical hacking and password auditing to wireless security and incident response, students embraced every challenge with curiosity, creativity, and enthusiasm. It was inspiring to see them develop both technical skills and the confidence to become responsible digital citizens and future cybersecurity leaders. Oh, and the meme contest gave them the opportunity to poke some fun at me too, which I enjoyed greatly!”

Empowering Future Scientists

All three camps were great successes, introducing K-12 students to the world of STEM and computing while imparting critical-thinking skills and the practical knowledge needed to navigate our technology-driven society. Outside of specific summer camps, the outreach and engagement team at RCAC has been hard at work to promote STEM and HPC to as many students as possible. The world needs future professionals who are excited by science and research and who can lead us into brighter tomorrows. To cultivate this love of STEM, RCAC has held workshops at the Children’s Museum of Indianapolis, delivered a hands-on 4-H club experience, and given multiple student-focused presentations throughout the nation. The center also directly enhances student and educational outcomes through its hands-on workforce development pipeline, the CI-XP (Cyber Infrastructure - eXPerience) Program.

“Talent is everywhere; access isn't,” says Suzanna Gardner, the Lead Research Operations Administrator for Outreach & Engagement at RCAC. “The people who change a field aren't usually the ones who had the most to start with; they're the ones someone took a chance on. That's what these camps do. Whether it's coding, cybersecurity, or building something for Mars, we're showing kids it's possible. And then through CI-XP, they get real work on real systems. But none of this happens without the team. The instructors, the CI-XP students helping run these camps, the mentors who show up. That community makes it real for students. That's what actually changes trajectories.”

All three summer camps will return in 2027. To learn more about RCAC’s outreach and engagement activities, please visit: RCAC Outreach and Engagement

RCAC would like to thank these partners and volunteers for contributing both their time and energy and making these camps possible: Nicholas Brasovan, George Bailey, Amanda Warren-Glowe, George Takahashi, Suzanna Gardner, Amari Gardner, Ava Sperazza, Bithi De, Tomas Ratkus, Luke Monroe, Taylor Graham, Samantha Lee Walker, Eric Adams, Michael Allan Carlson, Nannan Shan, Manasa Vemuri, Damieon Brown, Smriti Bhatt, Madiha Fathima, Banon Monroe, and Courtney Bragg.

To learn more about CI-XP opportunities, please visit: CI-XP Student Program

RCAC operates the centrally-maintained research computing resources at Purdue University, providing access to leading-edge computational and data storage systems as well as expertise and support to Purdue faculty, staff, and student researchers. To learn more about HPC and how RCAC can help you, please visit: https://www.rcac.purdue.edu/ or reach out to rcac-help@purdue.edu to request a consultation.

These camps were supported by the Anvil supercomputer. Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

Shubham Jha is a second-year Master's student in Engineering Management at Purdue University. He came to Purdue in August of 2024 after working as a full stack developer for Cisco. After settling in at the university, Jha decided to apply for a job at RCAC through the CI-XP program. Jha was initially hired in May of 2025 onto the Research Software Engineering (RSE) team as a front-end developer for RCAC’s proprietary unified HPC center operations portal, Halcyon. Since then, Jha has grown far beyond that original role: leading application development for the U.S. Air Force, architecting agentic AI solutions for cross-institutional collaborations, and designing the disaster-recovery strategy for a $4 million NSF project — the list is extensive. Again and again and again, Jha has proven himself an invaluable asset to RCAC’s team, but his experience with the CI-XP program is not a happy accident, it’s a design feature.

CI-XP prides itself on the fact that the program does not partake in “toy” projects. Students gain real workplace experience by taking on tasks with real-world impact. Interest in different pathways and fields is often rewarded; in Jha’s case, he began as a front-end website developer, found an interest in other areas, and was encouraged to pursue those interests:

“My experience with RCAC and the CI-XP program has been wonderful,” says Jha. “Learning-wise, when I was at Cisco, we would get a single project and our deliverables focused only on that. So when I was a network engineer, I only used to write Python scripts, or I only used to fix bugs. When I transitioned to the front-end engineering team, I only used to do front-end development. But at RCAC, in the span of one year, I have been able to touch different types of projects, initially starting as a front-end developer, then moving on to the back-end and eventually agentic AI. RCAC gives me the time and opportunity to ramp up, so I am able to learn about things from different researchers and I'm also able to apply those learnings practically.”

Jha continues, “Getting to work on different technologies, different aspects of what’s relevant in the industry, that’s what I enjoy most about RCAC. Whenever I want to grow or expand my knowledge base or my technical expertise, I just ask Daniel [Madren, Lead Research Development Administrator] and he finds a way to get me onto those projects. I can learn about different domains and actually work on them. It's one of the best things — I'm not restricted to a specific task for a specific responsibility, and I can explore different projects.”

Jha’s experience is not unique for the CI-XP program, but he more than earned the trust required from leads to put a student on such important projects.

"Shubham has a rare combination of technical ability, sound judgment, initiative, and interpersonal skill,” says Daniel Madren, Program Manager for the Research Software Engineering Center at RCAC and a CI-XP mentor. “He has earned the trust of our team by communicating clearly, collaborating across functions, and taking ownership of complex work without needing to be closely directed. He does not simply complete assigned tasks, he identifies what needs to be done and helps move the work forward. Over the past year, he has become a valued colleague and a true team player. I have learned as much from Shubham as he has learned through the CI-XP program, and working with him has been a genuine pleasure."

In his time thus far at RCAC, Jha has worked on nine distinct projects, contributing in multiple ways to each one. One notable contribution is Jha’s work on the StreamCI project. StreamCI is a real-time sensor data management and processing infrastructure hosted entirely on Purdue’s Anvil supercomputer. Funded by a $4 million grant via the NSF CSSI program, StreamCI is designed as an artificial intelligence (AI)-ready streaming data platform used to support diverse scientific applications. Jha’s role on the team was varied. He began by developing the initial landing page for the StreamCI website. Then, Jha helped transition the platform from an older streaming backbone to a newer one preferred by the team. Specifically, he engineered the migration from RabbitMQ to a production 3-node Apache Kafka cluster with replicated, partitioned topics and 30-day retention, enabling new processing stages to be added without modifying existing ingestion paths. After the Kafka migration, Jha set his sights on StreamCI’s data backup strategy.

Being able to perform disaster recovery is of the utmost importance for research teams who rely on massive volumes of data for their work. To ensure StreamCI had this capability, Jha built an end-to-end S3 backup infrastructure — from the initial S3 backup foundation to a unified Kustomize base — plus development/production overlays with standalone CronJobs, dual-endpoint S3 support (separate source/destination), and a restore-benchmark harness. He also added point-in-time recovery with a 6-hourly oplog incremental CronJob for production, a shared PITR marker across full/incremental jobs, and extended oplog retention so restores can roll forward between full backups. Simply put, Jha solved the data backup issue.

“What stood out about Shubham was the ownership he took — he didn't just close out tasks, he looked at where StreamCI was heading and solved the hard problems before they became blockers,” says Jaewoo Shin, a Research Scientist at RCAC and the technical lead for StreamCI. “The Kafka migration and the backup and disaster-recovery infrastructure he built are still load-bearing parts of the platform today, and he designed them so cleanly that our team could keep extending the system without ever having to untangle his work. As the dev lead, that's the highest compliment I can give an engineer — I could hand him a critical piece of the platform and never worry about it again.”

Another notable project that Jha worked on was the ACID-R (Automated Commercial Industry Data – Repository) platform. ACID-R is designed to help the U.S. Air Force identify promising new technologies with more efficiency by harnessing the power of AI, but with no risk of hallucination (AI-fabricated false information). The platform connects commercial vendors with the Air Force — vendors can upload their capability statements, and the Air Force can view, search, and filter through them all (thousands of proposals) with the help of AI. Jha assisted with this project by tackling:

• Front-end Dashboard development and design
• Back-end development with APIs (Application Programming Interface)
• Integration of the platform with RAGs (Retrieval-Augmented Generation)
• Deployment to Geddes (RCAC’s Community Composable Platform) and Rowdy (RCAC’s technology testbed)

ACID-R also happens to be Jha’s favorite project so far:

“I really enjoyed working on the Air Force project because at Cisco, I only worked as a front-end developer, in which I was given a Figma file by my team, and I just created a website based on the Figma file,” says Jha. “But on this project, we actually didn't have the design. We had to come up with one that would solve their problem. So we had to work with AI researchers like Elham [Barezi, Lead AI Research Scientist at RCAC] to actually achieve the functionality that would satisfy the customers. I learned about RAG pipelines and was able to deploy the application on Rowdy as well. Before this, I had no experience working on deployment. I would only run applications on my system and then give it to the DevOps team, and they would deploy it on other systems. Getting to do everything from the requirements gathering, coordinating with the customers and AI researchers, deploying it myself and optimizing it, and getting to learn while I was doing this project was a great experience for me, and that's why it's my favorite project so far at RCAC.”

The amount of work the Jha has accomplished throughout his time at RCAC is astonishing, especially considering that he is concurrently pursuing his Master’s. Thankfully, the CI-XP program understands that a student’s education comes first and allows for a lot of flexibility with work hours.

“Because Daniel is quite flexible, I am able to balance my work at RCAC based on my hours,” says Jha. “I can do the work whenever I have free time, instead of on a set schedule. That’s how I have been able to manage projects while still following my day-to-day Master’s schedule.”

The CI-XP program also encourages students to attend conferences in their field and even helps support them financially through travel scholarships. Because of his association with the program, Jha has been able to attend numerous conferences while at RCAC, including: SC25, PEARC25, Nvidia GTC 2026, DesignCon 2026, and Developer Week 2026. Thus far, he has liked SuperComputing (SC) the most, which is unsurprising given that it is the largest international conference for HPC.

“Supercomputing was my favorite one so far because it's the biggest conference I have ever attended, and seeing so many solutions on the floor, getting to interact with professionals involved in HPC — it was great,” says Jha. “I met product managers from Microsoft and HP. I got the opportunity to learn about quantum AI and saw how I can use different technologies in my own personal work. For instance, I took a workshop on Agentic AI which really helped me to upskill myself and understand what the Telecoupling Toolbox project [a Michigan State University collaboration Jha is currently working on] actually requires, and come up with the architecture for it.”

In his time at RCAC, Jha has made a huge impact on the organization and the various projects he has been a part of, not only writing software, but shaping what gets built and why. Even more importantly, he has enjoyed the CI-XP program and benefited tremendously from it, which is why he highly recommends it to other students at Purdue:

“My experience with RCAC has been wonderful,” says Jha. “I’ve gotten to learn a lot, network with a lot of people, and attend some great conferences. I would definitely recommend looking for opportunities with the CI-XP program. You get to learn from the researchers at RCAC, and all of them are so helpful. Whenever I have issues or need some clarification, I can just reach out and all of them are ready to help me. The team is also really understanding regarding your other commitments, and makes accommodations for you. The program helps you in the long run as well by building up your profile and resume. It’s great.”

To learn more about CI-XP opportunities, please visit: CI-XP Student Program

The Purdue Center for Research Software Engineering (aka the RSE center) is a university-approved center within RCAC. Its official establishment recognizes the increasingly vital role that software plays in all fields of scientific research, and formalizes RCAC’s software engineering efforts at RCAC to better support research at Purdue. The RSE center’s mission is to help accelerate research and increase its impact through the creation of innovative, robust, and sustainable research software. To learn more about the center and how we can help you, please visit: Purdue RSE Center

To learn more about High-Performance Computing and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

The regional NAIRR AI workshop was a two-day event held in Indianapolis, Indiana. Day one opened with breakfast in the atrium, followed by a welcome from Marisa Brazil, Director of Strategic Initiatives in the Research Technology Office at ASU, and opening remarks from Preston Smith, Executive Director of RCAC. Daniel Schiff then took to the stage and delivered the keynote presentation for the event. Schiff is an Assistant Professor of Technology Policy at Purdue’s Department of Political Science and the Co-Director of the Governance and Responsible AI Lab (GRAIL). His presentation was titled “Turn off that Chatbot! Conjectures on Adapting to AI in the Classroom,” which covered the multiple considerations and quagmires that professors (and students) need to navigate in today’s AI-enabled educational landscape. Schiff’s keynote led to a lively and productive discussion amongst the attendees and was a great way to start off the workshop.

Following the keynote, the first day was filled with insightful presentations and work sessions that introduced attendees to new advancements in AI and machine learning (ML) applications for HPC systems. The list of sessions was as follows:

• The Agentic Era: AI Operators and the New Responsibilities of Research, Geoffrey Lentner, Principal AI Scientist, RCAC, Purdue University
• AI Modalities and Research Tools: From LLMs to Domain-Specific Models, Juanjo Garcia Mesa, Research Software Engineer, Arizona State University
• The New Laboratory: Building Trustworthy AI for the Future of Science and Engineering, Guang Lin, Professor of Mathematics, Purdue University
• Real Applications of Machine Learning (REALM), Bharat Bhargava, Professor of Computer Science, Purdue University
• Towards Safe and Robust Vision Models, Raymond Yeh, Assistant Professor of Computer Science, Purdue University

To cap off the day, the event organizers held an Open Office Hours session, in which attendees could participate in informal discussions, engage in technical Q&A’s, and receive hands-on guidance from workshop facilitators and national cyberinfrastructure experts. Workshop attendees were encouraged to bring their own computational workflows or challenges to the session so they could ask specific questions and receive assistance for their computing issues. Multiple AI and ML HPC experts were on-site to offer support, including representatives for Purdue’s Anvil system and Indiana University’s Jetstream 2, as well as facilitators from NAIRR.

Day two began with a welcome and run-of-show delivered once more by Brazil, followed by a special presentation from David Crandall. Crandall is a professor of Computer Science and the Director of the Luddy Artificial Intelligence Center at Indiana University (IU). His presentation, titled “Egocentric Computer Vision, for Fun and for Science,” focused on first-person computer vision research. Most computer vision work has been conducted from a third-person perspective—images and videos from surveillance cameras and smartphones. But Crandal’s research concentrates on social relationships between people by utilizing lightweight cameras that approximate a true human field of view. Crandal argues that this “egocentric” perspective better captures social interactions and will therefore lead to breakthroughs not only in computer vision, but in psychological and developmental research. During his presentation, he shared the pros and cons of research from these two perspectives, as well as the progress and results of various projects conducted at his lab.

The next session for the day was a panel discussion titled “The Indiana AI Corridor: How R1 Research Computing Infrastructure is Democratizing AI for Every Indiana Researcher.” This discussion touched on the importance of research computing infrastructure and how investments at universities benefit researchers throughout the state. Notably, the group discussed how these investments are only maximized if researchers at Indiana's smaller colleges and regional institutions can access, navigate, and effectively use the resources. The conversation examined the importance of human-driven support and training, and highlighted how the strategic Federal and institutional investments already in place at Indiana’s three R1 universities all converge to create a consortium for access to advanced AI computing resources, making AI research infrastructure available to all. The panelists included: Eric Adams, Lead Research Operations Administrator for Education at RCAC; Scott Michael, Director of Research Software and Solutions at IU; Caleb Reinking, Associate Director of Research Software Engineering at Notre Dame University; and moderator Gil Speyer, Director of the Computational Research Accelerator at ASU.

After the panel discussion, the rest of Day two proceeded as follows:

• Discovering and Controlling Safety Risks in Foundation Models: A Probabilistic Perspective, Ruqi Zhang, Assistant Professor of Computer Science, Purdue University
• Why Most Scientific Data is Not AI-Ready (and How to Fix It), Nathan Denny, Lead Research Analyst, Purdue University
• Overview of National Cyberinfrastructure Resources, Eric Adams, Lead Research Operations Administrator, RCAC, Purdue University, and Juanjo Garcia Mesa, Research Software Engineer, Arizona State University

The two-day event came to a close with final remarks given by Brazil and another Open Office Hours session. To view the full schedule and see a brief synopsis of each presentation, please visit: NSF NAIRR Regional AI Workshop

RCAC operates the centrally-maintained research computing resources at Purdue University, providing access to leading-edge computational and data storage systems as well as expertise and support to Purdue faculty, staff, and student researchers. To learn more about HPC and how RCAC can help you, please visit: https://www.rcac.purdue.edu/ or reach out to rcac-help@purdue.edu to request consultation.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot. Indiana researchers, educators, and businesses can leverage the Indiana Consortium for AI InnovatioN (I-CAN) to for onboarding to advanced computing resources like Anvil and Indiana’s Jetstream 2.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

Dr. Zezhou Cheng is an Assistant Professor of Computer Science at the University of Virginia, where he leads the Computer Vision Lab. He and his graduate students have taken on a slew of research projects involving an advanced computer vision technique called 3D object detection. Traditional 2D object detection systems can locate and detect objects on flat images by analyzing the object’s horizontal and vertical properties. A major limitation, however, is a lack of depth data, meaning these systems cannot tell how big an object is, how far away it may be, or how it is oriented in space. 3D object detection systems overcome this problem by analyzing point clouds, which are created with data obtained from sensors such as LiDAR and stereo cameras. This depth data is crucial for applications such as self-driving vehicles, robotics, and augmented reality systems, all of which need precise spatial awareness to function properly.

Cheng’s group has been quite prolific in the computer vision research space over the past two years—five publications in 2025, five so far in 2026, as well as a preprint currently awaiting approval. Their output is impressive, but conducting AI-driven research at such a pace requires access to cutting-edge high-performance computing (HPC) systems. For this, Cheng turned to the Anvil supercomputer.

Cheng applied for an allocation on Anvil through the National Artificial Intelligence Research Resource (NAIRR) Pilot program. Once approved, Cheng and his students had access to multiple NVIDIA H100 GPUs, allowing them to train and develop AI models at a rate unimaginable on traditional computing platforms.

“Anvil is wonderful. One of the best, I would say,” says Cheng. “There are many H100 GPUs available. Students could schedule jobs on 12 to 14 H100s, which enabled us to train very large models. A single GPU is too limited. On other systems, we would have to wait up to three days for 8 A100s [NVIDIA’s older GPU chip], but on Anvil, we could launch 8 H100s within a few hours.”

One particularly impactful project was LabelAny3D, an automatic 3D annotation pipeline that efficiently generates 3D bounding boxes for objects across arbitrary categories through a technique known as monocular 3D object detection. This technique relies not on point clouds and depth data collected from sensors, but from a single RGB image. Monocular methods are more efficient and accessible than approaches that require specialized sensors, and researchers believe they are key to deriving real-world impact from 3D object detection systems.

Cheng and his group developed LabelAny3D as a way to overcome a major challenge in AI model training—annotation. Annotation refers to labeling individual images used within AI model training datasets. Typically, this is a time and labor-intensive process that needs to be conducted by a human. Current monocular 3D detection models were designed with autonomous driving in mind, so the people annotating the datasets focused on cars and pedestrians at the expense of anything else. This inevitably leads to models that struggle with images “in-the-wild.”

“Because it is useful for autonomous driving, people have the motivation to annotate vehicles and pedestrians,” says Cheng. “That’s why current 3D datasets contain limited categories, not including categories such as animals or food. LabelAny3D is designed to fill this gap and unlock 3D detection in-the-wild.”

With LabelAny3D, Cheng and team have developed a framework that automatically reconstructs 3D representations from 2D images and produces high-quality 3D bounding box annotations. Using LabelAny3D, the group then developed a new benchmark for open-vocabulary monocular 3D detection that includes over 80 object categories, most of which are absent from existing 3D datasets. Future researchers can now use Cheng’s 3D annotation pipeline to create higher-quality AI training datasets as well as more robust and functional 3D recognition systems for real-world applications.

Cheng presented LabelAny3D at the Conference on Neural Information Processing Systems (NeurIPS) 2025, a renowned international machine learning and computational neuroscience conference. While there, Cheng also presented another research project enabled by Anvil: Frame In-N-Out. He has since gone on to present other Anvil-supported work at various conferences, including the International Conference on 3D Vision (3DV) 2026, the International Conference on Learning Representations (ICLR) 2026, the Computer Vision and Pattern Recognition (CVPR) Conference 2026, and the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2026. Needless to say, this level of high-quality AI-driven work would not be possible without a powerful HPC resource like Anvil.

Another great aspect of having access to the Anvil supercomputer is the robust support team that comes with it. Cheng was thrilled with the level of support he received from the Anvil team.

“The Anvil team,” says Cheng, “they really spent a lot of time helping me solve some storage limitation issues, and we really appreciate their help. One problem was that we needed to download a very large dataset—probably 30TB or more—and the team offered to download it for us and provide it on the system in a shared space so future users could access it as well.”

Cheng continues, “They also provided same-day feedback for questions like adding new users or increasing SUs. They were very helpful.”

For more information on Cheng and the Computer Vision Lab’s research, please visit: Publications

To learn more about High-Performance Computing and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Publications citing Anvil:

1. Wentao Zhou, Xuweiyi Chen, Vignesh Rajagopal, Jeffrey Chen, Rohan Chandra, Zezhou Cheng. Stereo4DWalker: Learning 4D-aware Embodied Urban Navigation from Internet Stereo Videos. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026
   
2. Xuweiyi Chen, Wentao Zhou, and Zezhou Cheng. Wildrayzer: Self-supervised large view synthesis in dynamic environments. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.
   
3. Xuweiyi Chen, Wentao Zhou, Aruni RoyChowdhury, and Zezhou Cheng. Point-moe: Towards cross-domain generalization in 3d semantic segmentation via mixture-of-experts. In International Conference on Learning Representations (ICLR), 2026.
   
4. Xuweiyi Chen and Zezhou Cheng. Semantic-free procedural 3d shapes are surprisingly good teachers. In International Conference on 3D Vision (3DV), 2026.
   
5. Jin Yao, Hao Gu, Xuweiyi Chen, Jiayun Wang, and Zezhou Cheng. Open vocabulary monocular 3d object detection. In International Conference on 3D Vision (3DV), 2026.
   
6. Jin Yao, Radowan Mahmud Redoy, Sebastian Elbaum, Matthew B. Dwyer, and Zezhou Cheng. Labelany3d: Label any object 3d in the wild. In Neural Information Processing Systems (NeurIPS), 2025.
   
7. Boyang Wang, Xuweiyi Chen, Matheus Gadelha, and Zezhou Cheng. Frame in-n-out: Unbounded controllable image-to-video generation. In NeurIPS (NeurIPS), 2025.

Written by: Jonathan Poole, poole43@purdue.edu

Object storage has exploded in popularity in recent years due to its ability to deliver virtually unlimited storage scalability. Unlike traditional file- or block-based storage, object storage treats each individual piece of data as a self-contained object. Every object contains three things—the data itself, the metadata, and a universally unique identifier—and is stored in a structurally flat container known as a “bucket.” Any object within the bucket can be quickly retrieved and analyzed (regardless of file type) based on the custom metadata. Thanks to this unique architecture, object storage allows users to store and manage extremely high volumes of unstructured data while retaining speed and data availability.

“Traditional filesystems can face significant performance bottlenecks when scaling to tens or hundreds of millions of files,” says Erik Gough, Senior Research Scientist at RCAC. “The performance impacts can be felt by all users of a shared filesystem when a single user is performing millions of metadata operations simultaneously. Object storage is designed specifically to eliminate this contention on filesystem metadata servers at large scale.”

Object storage is ideal for researchers working with large volumes of static data, such as those leveraging artificial intelligence (AI) and machine learning (ML) methods in their work. Typical use-cases for object storage include creating and utilizing scientific data lakes, instantly accessing ML/AI training datasets, and archiving large volumes of rich media content. With Anvil Object Storage, researchers now have access to a storage pool that can scale to petabyte levels without the metadata performance bottlenecks typically found in traditional file systems, allowing for the storage of billions of items without performance degradation.

“With Anvil Object Storage,” says Gough, “we are giving Purdue and national researchers a powerful new storage tier that is optimized for large-scale datasets that are required for AI and LLM training. We’re excited to see how this new storage tier impacts scientific discovery on Anvil.”

Anvil Object Storage is a 1-petabyte (PB) high-performance, software-defined storage tier integrated directly into the Anvil Composable Subsystem. The new storage service is powered by Ceph, providing scalable, all-flash S3-compatible storage to Anvil users. Researchers can create dedicated storage buckets to securely store, access, and manage their data using familiar S3 client tools such as rclone or s3cmd. The service provides a REST-based API endpoint accessible from Anvil systems and supports flexible access controls for modern data-intensive workflows.

Anvil Object Storage is specifically optimized to host the massive image, text, and sensor datasets required for training Large Language Models (LLMs) and complex computer vision systems. The storage is directly accessible from Anvil’s nodes (including NVIDIA A100 and H100 nodes), facilitating rapid data ingestion for AI training and model hosting.

Access is currently available upon request, subject to available capacity, with a total system capacity of 1PB. To request access, please submit a ticket through the ACCESS or NAIRR ticketing system.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

RCAC’s core mission is to provide access to leading-edge computational and data storage systems as well as expertise in a broad range of HPC activities. Outside of simply being a resource and expertise provider, the center also strives to actively promote the effective use of HPC and to lead the charge in developing the HPC workforce of tomorrow. RCAC has a long history of engaging with students to stoke interest in HPC. The CI-XP Program, for instance, teaches crucial computing skills that will be needed after graduation, and provides real workplace opportunities so that they may grow as professionals in the field. With this in mind, the outreach and engagement team decided to sponsor two amazing, student-focused hackathons this year.

“I spent years coaching athletes at the college, national, and international level, so high performance has always been my thing,” says Suzanna Gardner, Lead Research Operations Administrator for Outreach & Engagement at RCAC. “Figuring out how to help people grow stronger, get more strategic, and build something that actually lasts. That mindset doesn't go away. It just finds new places to show up. For us, one of those places is hackathons. We sponsor them because access matters. Students are among the most talented, driven people out there, and many of them just haven't had a seat at the table yet. When RCAC shows up at a hackathon, we're not there to put our logo on a banner and call it a day. We're there to say: this supercomputer, this technology, this future - it belongs to you too. That's how you build the next generation of researchers. Not by waiting for them to find their way in, but by showing up and opening the door.”

The first hackathon of the semester was Hack Indy 2026. This event is the largest student-run hackathon at Purdue University in Indianapolis, with over 200 students participating. Hack Indy took place over the course of three days. Students faced the challenge of building any software project they wanted, with the caveat that it must be built entirely during the hackathon—no prior development was allowed. Different prizes were offered for specific categories, such as “Best use of Gemini API” or “Best use of ElevnLabs.” RCAC sponsored Hack Indy 2026 by providing computing resources on the NSF-funded Anvil supercomputer. With this sponsorship, the participants had access to the Anvil-AI partition, which was recently upgraded to include 84 Nvidia 80GB H100 SXM GPUs, meaning Hack Indy users could make full use of highly sought after, cutting-edge compute resources typically reserved for advanced research.

Outside of competing, Hack Indy also provided participants with a weekend filled with keynote presentations and workshops given by industry leaders. As part of this, RCAC hosted a one-hour hands-on workshop on supercomputing, teaching students the basics of HPC and highlighting different ways computing can be used for science. RCAC staff members were also on hand through the entirety of the weekend to answer questions or assist with any issues.

The second hackathon of the semester was Catapult 2026. Catapult 2026 was a 36-hour AI + ML × Entrepreneurship hackathon hosted by ML@Purdue, a student organization for machine learning and artificial intelligence. The event was designed for students to come together, build innovative projects, attend workshops, and compete for prizes. RCAC offered 1000 SUs on the Gautschi-AI partition for students to use during the hackathon. Gautschi is Purdue’s most powerful supercomputer, with Gautschi-AI boasting a whopping 160 next-generation NVIDIA H100 GPUs. Thanks to having Gautschi access, the students could tackle any project they wanted during the hackathon without experiencing any computing slow down or lag time.

For Catapult 2026, RCAC also hosted a booth at the event, provided on-site expertise, gave presentations on HPC, and offered a prize for “Best Machine Learning Project.” The winning team developed VIE. VIE is a web tool that combats AI facial recognition systems by making faces in photos unidentifiable to computer systems, but without any perceptible change to human eyes. The group relied heavily on the Gautschi H100 GPUs to develop VIE. For their prize, RCAC will provide the team with 750 SUs on Gautschi-AI, which they can utilize over the course of one year.

“We've been talking with the ML@Purdue student organization for a few years now," says Geoffrey Lentner, Principal AI Scientists at RCAC. "It's been a privilege to work with them facilitating practice and experience in HPC for student projects. The Catapult hackathon is always a fun experience - I enjoy discussing data science and AI concepts and new ideas with the students. I'm impressed with the skills some of the undergraduates already poses and ML@Purdue is a big part of giving some of these students the chance to build these skills. For some students the hackathon is their first exposure to resources like this, and that's great too.”

In continuing with its mission to develop the HPC workforce of the future, RCAC will be hosting multiple summer camps this year. These camps are aimed at middle and high school students looking to gain hands-on experience with computing, STEM, and HPC.

The first camp is “Code Explorers: Coding and AI.” Starting on June 21, this one-week camp will help students discover the power of coding, data science, and environmental action. Participants will get a taste of the full research pipeline: from collecting real-world environmental data to analyzing that data using essential coding skills to presenting their results in a compelling and visually satisfying way. The students will also explore the fundamentals of machine learning, write their own Python code, and learn to refine their code using AI techniques. To learn more, please visit: Code Explorers: Coding and AI (registration closed due to max capacity)

The second camp is another week-long immersive experience, called “CyberSafe Heroes: Empowering High School Students for a Secure Digital Future.” This camp begins on July 12 and is designed to spark interest in cybersecurity and equip students with essential digital skills through hands-on activities. Aligned with national cybersecurity education standards, the program covers encryption, ethical hacking, and online safety while promoting responsible technology use and digital citizenship. The students will engage in encryption challenges, ethical hacking simulations, and cybersecurity escape rooms, explore how artificial intelligence (AI) is used in modern cybersecurity, and even connect with industry professionals through career panels. To learn more, please visit: CyberSafe Heroes: Empowering High School Students for a Secure Digital Future (registration closed due to max capacity)

The final summer camp offering from RCAC is the Mission to Mars Summer Day Camp. This camp is designed for incoming 7th-9th grade students. Participants will step into the role of Mission Control during a simulated Mars mission in crisis. Through three interconnected, story-driven STEM challenges, campers will investigate real problems, design creative solutions, and collaborate with peers - all in a race to save the mission. Three one-day sessions are currently available:

• Session 1 - June 15, 2026
• Session 2 - June 16, 2026
• Session 3 - June 17, 2026

Each session runs from 9:00 AM – 4:00 PM. Snacks and lunch are included. To learn more about the day camp and to register, please visit: Mission to Mars Summer Day Camp

RCAC operates the centrally-maintained research computing resources at Purdue University, providing access to leading-edge computational and data storage systems as well as expertise and support to Purdue faculty, staff, and student researchers. To learn more about HPC and how RCAC can help you, please visit: https://www.rcac.purdue.edu/ or reach out to rcac-help@purdue.edu to request consultation.

Written by: Jonathan Poole, poole43@purdue.edu

Datasets are invaluable for research, but this holds especially true for AI research. AI modeling and machine learning typically rely on immediate access to massive collections of data, whether the end goal is to train a new model or to use an existing one for more specific research. Domain-specific datasets accommodate a researcher’s needs via a single package. However, these packets of data have a downside.

The problem inherent in datasets is that they contain exactly what a researcher wants—a colossal amount of data. The sheer volume of information contained within datasets entails an extraordinary storage footprint, long transfer time, and impact on the machine’s memory. Even researchers utilizing HPC resources can be waylaid by the effort of obtaining the datasets they need and ensuring they are located where their system can use them. The Anvil team at the Rosen Center for Advanced Computing (RCAC) decided to help researchers bypass this issue by amassing datasets into a data repository that is pre-downloaded and ready for use on an HPC system, backed by a fast underlying network. Now, anyone with access to Anvil can use these datasets immediately in their work, saving them both time and hassle on their projects, and accelerating research.

“Making popular datasets natively available on Anvil fundamentally changes how researchers work,” said Haniye Kashgarani. “Datasets with very large numbers of files are hosted in Anvil Object Storage and are also made available in optimized formats such as SquashFS and LMDB on the Anvil file system. This immediate, high-performance access allows scientists to fully leverage HPC and AI workflows without the overhead of data transfers, storage constraints, or redundant downloads. As demand grows, additional widely used datasets can be added to the platform upon request.”

The most recent additions to Anvil’s data repositories are its AI datasets. This collection covers computer vision, PhysicalAI, and robotics, and supports tasks such as detection, segmentation, tracking, control, reinforcement learning, and large-scale model pretraining and evaluation across domains, including everyday objects, smart spaces, and embodied PhysicalAI. There are currently nine datasets in the collection, with more to come. The new AI datasets will enable scientists on Anvil to leverage machine learning techniques and quickly develop AI models that can be embedded into physical systems such as robots or drones, without needing to download and manage the data themselves.

In addition to the new AI datasets, Anvil hosts dataset collections for geospatial, hydrological, meteorological, covariates, igenomes, and GeoAI research. In total, these collections amount to over 215TB worth of data. RCAC’s efforts to centralize and host these valuable datasets on Anvil make the data more easily discoverable, accessible, and usable for scientists throughout the nation.

“One of our goals with Anvil is to push the limits of scientific discovery,” says Arman Pazouki, Director of Scientific Applications at RCAC and co-PI on the Anvil project. “Hosting these datasets on Anvil makes research more efficient, allowing our users to focus on conducting science instead of on data management. As a result, researchers will be able to harness the power of AI and machine learning easier than ever before, expediting the rate at which scientific breakthroughs are possible. This is just one small step in how Anvil is helping to reshape the world of research and expand access on a national scale.”

Researchers who would like to use Anvil’s datasets can learn more here: Anvil Dataset Documentation

To learn more about High-Performance Computing and how it can help you, please visit our “Why HPC?” page.

Anvil is one of Purdue University’s most powerful supercomputers, providing researchers from diverse backgrounds with advanced computing capabilities. Built through a $10 million system acquisition grant from the National Science Foundation (NSF), Anvil supports scientific discovery by providing resources through the NSF’s Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS), a program that serves tens of thousands of researchers across the United States. Anvil also supports advanced artificial intelligence research as an official resource provider of the National Artificial Intelligence Research Resource (NAIRR) Pilot.

Researchers may request access to Anvil via the ACCESS allocations process or through the NAIRR allocations process. More information about Anvil is available on Purdue’s Anvil website. Anyone with questions should contact anvil@purdue.edu. Anvil is funded under NSF award No. 2005632.

Written by: Jonathan Poole, poole43@purdue.edu

We will provide an update by noon, April 28 or sooner as work progresses.