Twenty Years of the Student Cluster Competition

From six teams in Reno to a worldwide proving ground for HPC talent, the Student Cluster Competition marks its 20th match-up this fall at SC26 in Chicago.

In November 2007, six teams of undergraduates built clusters at SC07 in Reno, Nevada. At the time, high-performance computing was rarely part of an undergraduate education. The competition was created to change that.

This November, the Student Cluster Competition (SCC) turns 20. Ten teams will compete on the SC26 show floor in Chicago, 15–20 November, and another 29 teams from 10 countries will compete remotely in SCC Connect. Attendees will also find a space on the exhibit floor dedicated to the milestone.

Two decades of success make for an impressive timeline. But the program’s true legacy is measured by what it changed: the courses universities now teach, the students from outside computer science who discovered a place in the field, and the alumni now shaping the future of computing at national laboratories, in industry, and in graduate programs.

Brent Gorda and Ricky Kendall created the competition with two goals: get students excited about HPC and push universities to teach it. Hosted by the Students@SC program, it has drawn undergraduate teams from around the world ever since.

The need was real. A 2010 IDC study found that 93% of HPC centers had trouble hiring staff with the right skills. Students in physics, biology, or chemistry rarely encountered parallel computing at all, and most computer science programs offered only a single course.

Six teams answered the first call: Indiana University, National Tsing Hua University, Purdue University, Stony Brook University, the University of Alberta, and the University of Colorado.

The format works because it mirrors the job. Teams partner with a vendor to design a cluster, learn a set of scientific applications, and tune each for the other. Then, they run a real workload under a fixed power limit while explaining their choices to judges and passing attendees.

A 2015 study by SCC organizers compared the competition to the procurement and acceptance testing of a professional HPC center. No one student can do it all, so teams learn to divide the work, trust one another, and communicate across disciplines.

Over time, organizers have made that experience even more realistic with surprise applications, reproducibility challenges, unexpected power loss, cloud resources, and artificial intelligence benchmarks. Each one pushes students to adapt, recover, and solve problems under pressure.

The applications give that work purpose, showing students that HPC is not just about faster machines. It can help model hurricanes, accelerate drug discovery, and tackle problems that matter far beyond the competition floor.

A Path Into the HPC Community

In a 2013 survey of former participants, every respondent said many of their teammates had gone on to graduate study or careers in HPC. More than half said HPC education was uncommon at their university, meaning the competition introduced them to a field they might not otherwise have found. One longtime advisor estimated that about 30% of their students ultimately pursue HPC careers.

“This event guides them into the HPC community,” another advisor told organizers.

At SC, that community becomes real. Students attend technical sessions, meet mentors and employers, present their work, and build professional networks. Many later return to the conference as researchers and professionals.

A Stronger Path into the Workforce

What began as an extracurricular effort at many universities has helped inspire formal HPC coursework. The skills required to compete gave faculty a practical foundation for preparing students for the workforce.

A Wider Doorway Into the Field

From 2010 to 2015, about one in six applicants came from majors outside computing. For many science students, SCC offered a first encounter with HPC while putting computing and domain-science students on the same team.

A Global Footprint

In 2010, 75% of team proposals came from U.S. institutions. By 2015, fewer than half did. Other organizations have since launched student cluster competitions modeled on the same idea.

The 2015 study was candid about who the competition was not yet reaching. Funding and floor space limited the field to 12 teams, and schools with fewer than 5,000 undergraduates were absent.

Since then, the competition has expanded who can participate.

When the pandemic moved SC20 online, 19 teams competed remotely. The following year, IndySCC launched as a 48-hour virtual competition designed to give newer teams a place to start. It grew from five teams in 2021 to 33 teams from 13 countries in 2025, when its overall winner was recognized as part of the official SC Awards.

The in-person competition has grown more inclusive as well. William Henry Harrison High School fielded a team in 2017, and Laney College became the first community college to compete in 2018. This year, IndySCC became SCC Connect, a name that reflects what the program has increasingly become: a way to connect more students, institutions, and countries with the HPC community.

Team High Performance Commenters

Clemson University

United States

This year, Clemson University teamed up with Coastal Carolina University, uniting students from diverse backgrounds, academic majors, and skill sets. Team members use Clemson’s Palmetto 2 Cluster, configuring their software stack to be nearly identical to it. Their cluster, Olympus, contains a head and five compute nodes (3 CPU; 2 GPU). To prepare for SCC at SC26, team members have been meeting twice a week to strategize and provide updates about application optimization. All team members have built Raspberry Pi clusters and have competed in separate hackathons and cluster competitions to improve their overall teamwork. The team’s collective course work includes parallel architecture, networking, and operating systems. 

Team NTU

National Taiwan University

Taiwan

Team NTU members represent the Department of Computer Science and Information Engineering at National Taiwan University. Although Team NTU captured last year’s SCC title, this team is all new. They united in late 2025 during Professor Chun-Yi Lee’s (the team’s current advisor) parallel programming course through their shared curiosity about high-performance computing. Team NTU’s new roster pairs experience in parallel programming and systems administration with a research-driven mindset. The team has built and tuned clusters, optimized scientific codes, and studied various architectures. This year, the “new” Team NTU intends to turn that foundation into faster, reproducible results on more HPC workloads. 

TEAM SegFAUlt

Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)

Germany

Although they come from different scientific backgrounds, the students on Team SegFAUlt united to continue a long tradition of FAU participation in cluster competitions dating back to 2013. Building on experience passed on by their peers, the team is confident in its ability to meet the challenges ahead. The team originally consisted of 10 students before they competed for a coveted spot on FAU’s core six-member SC26 team. While none have competed at the SC conference before, each person has unique competitive skills beneficial for different tasks, both from a social and technical perspective, that will be especially valuable during their first-ever SCC@SC competition.

Nikita Bloshchanevich, Anna-Lena Burger, Oliver Kilo, Stephane Montcheu Tchuatui, Faris Muzaferija, Leon Weckert, and Jan Laukemann

Nikita Bloshchanevich

Leon Weckert

Jan Laukemann

(Advisor)

TEAM ISAMBARDS

University of Bristol

United Kingdom

The Isambards team features six competitors with two reserves from five universities spanning the United Kingdom. The Bristol Centre for Supercomputing (BriCS), which operates the UK’s fastest supercomputer, Isambard-AI, supports the team with training resources and guidance from its researchers and system administrators. These materials can be accessed from anywhere in the UK. HPE sponsors the Isambards’ competition hardware. For the team, preparation is their edge: every benchmark planned and tuned on production hardware is completed at different universities by team members with various computer science backgrounds, such as networking, machine learning, and system administration. 

TEAM NTHU SCC

National Tsing Hua University (NTHU)

Taiwan

Continuing its tradition of solid performance in previous years, NTHU SCC returns to SCC competition at SC26. This year’s team unites two competitors with prior SC experience and four first-time participants selected from NTHU’s student cluster competition program based on their outstanding performance. Collectively, the members have competed in prestigious international competitions, affording them a well-rounded foundation of competitive HPC experience. The team’s training is built around NTHU’s Advanced High Performance Computing Cluster Practice course, combining Linux administration, cluster deployment, parallel programming, and performance optimization. Former SCC competitors mentor new members with additional support from NVIDIA, AMD, and Gigabyte, providing industry insights and technical guidance. The team also prepares on flagship HPC systems at Taiwan’s National Center for High-performance Computing, gaining hands-on experience under real-world conditions.

TEAM SUPERNOVA

Nanyang Technological University

Singapore

Team Supernova unites students with diverse high-performance computing expertise. Some members are completely new to HPC competitions, while others have participated in regional contests. Team Supernova came together through the university’s HPC community, united by a shared interest in HPC, artificial intelligence, and systems optimization. Preparation for SC26 combines structured training with hands-on optimization and collaboration. Throughout the year, the team meets regularly with experienced mentors to work through HPC fundamentals, Linux system administration, parallel programming, and cluster management. Their training is supported by Singapore’s National Supercomputing Centre, which provides access to HPC resources and technical guidance. The team also works closely with industry technologies and software ecosystems from partners such as NVIDIA, AMD, and Supermicro, gaining hands-on experience with the hardware and tools commonly used in modern HPC environments. The team is excited for the upcoming SCC challenge at SC26 and looks forward to exchanging ideas and learning from the broader HPC community.

TEAM SEA SHELL

San Diego Supercomputer Center (SDSC)

United States

With two returning and four new members, Team Sea Shell unites students across diverse majors and class levels to explore and cultivate an interest in high-performance computing. Every member has gained HPC experience through UCSD Supercomputing Club projects and the Single Board Cluster Competition (SBCC), which was started by UCSD SCC alumni and has continued for the past four years. In SBCC, students build and administer a cluster of single-board computers then run benchmarks and scientific applications. Team members support each other during projects, classes, and competitions to acquire computing hardware and software skillsets. In addition to SBCC, Team Sea Shell also contributes to a broader campus community and is known for hands-on technical workshops and collaborative engineering projects. With mentorship from supportive SDSC and UCSD faculty and guidance from competition alumni, they prepare extensively on SDSC’s Expanse system, club clusters, and vendor research clusters.

TEAM RACKlette

Swiss Federal Institute of Technology Zürich

Switzerland

Founded in 2018 and entirely student driven, Team RACKlette is returning to SCC competition at SC26 with a team built around an intentional mix of new and experienced members. New members are paired with experienced ones from the start, so the applications, tools, and debugging approaches from past competitions get shared across the group. It also means newcomers get up to speed faster because they can lean on teammates who understand the process. Team RACKlette includes students from computer science, computational science and engineering, and physics. Moreover, prior HPC experience is not a requirement to join the team. They prioritize motivation and a willingness to learn. As such, the team is much larger than the six SCC competitors because non-competing members regularly join training sessions and internal workshops, maintaining a steady pipeline of people learning HPC who are ready to step up in future years.

TEAM GIG ‘EM BYTES

Texas A&M University

United States

Gig ’Em Bytes is back for another run at SCC, and, this time, the team is larger than ever. This diverse, multidisciplinary team proudly runs out of Texas A&M’s High Performance Research Computing (HPRC) center and pulls its six competitors from four distinct departments: computer science, computer engineering, aerospace engineering, and mathematics. Each member leads an application area and shares expertise with the team, drawing on experience from internships, research, HPRC systems, and hands-on high-performance computing work. Behind the scenes, a second group of students trains alongside them, learning more about HPC and serving as a future pipeline for the team. Gig ’Em Bytes has been preparing since late spring and through the summer, starting with biweekly check-ins to review progress and build up HPC fundamentals. The workload is split into small groups of two with each pair specializing in a single benchmark or application. Gig ’Em Bytes is grateful to Texas A&M’s HPRC center for providing the team with learning resources, access to real-world supercomputers for team development, and faculty support.

TEAM PHOENIX

Georgia Institute of Technology

United States

SC26 marks Georgia Tech’s return to the SCC after its last appearance in 2021. Since then, Team Phoenix has competed in SCC Connect (formerly IndySCC), including in-person participation at SC24 and SC25. At SC25, Team Phoenix placed first among 11 U.S. teams and sixth among 30 teams overall. For SCC@SC26, the team will field four newcomers whose experience spans systems administration, cloud infrastructure, GPU orchestration, observability, and open-source education. Team Phoenix is part of the AI in the Supercomputing Era (AI-ESC) Vertically Integrated Projects, a course that started in spring 2020 with support from Georgia Tech’s College of Computing and Partnership for an Advanced Computing Environment (PACE). In VIP, students develop practical HPC skills through progressive projects in containerization, resource management, software and driver configuration, application benchmarking, and cluster analytics across diverse architectures. In preparation for SCC, Team Phoenix team also learned how clusters are designed, procured, and operated through vendor discussions, data-center tours, and hands-on server dissection. This preparation has led to student-developed tools deployed on Georgia Tech’s PACE clusters, novel-architecture testbeds, and the AI Makerspace.