EPFL donates Helvetios supercomputer to Universidad Nacional de Córdoba, Argentina; will triple UNC's computing capacity

EPFL donates Helvetios supercomputer to Universidad Nacional de Córdoba, Argentina; will triple UNC’s computing capacity

Helvetios Supercomputer begins a Second Life in Latin America - EPFL

Helvetios Supercomputer begins a Second Life in Latin America

Import & publish the news

Are you sure you want to import this news into ?

This news will be sent to its subscribers.

As part of its sustainability strategy, EPFL maximizes the use of its computing infrastructure throughout its lifecycle, helping reduce the environmental impact of the IT sector. The donation of the Helvetios cluster to the Universidad Nacional de Córdoba (UNC) in Argentina is a concrete example of this approach, giving the system a second life.

This cluster represents a massive surge in computing power for our institution, in scale, storage, and networking capabilities. After just one day, the mindset of our technical team has shifted completely: they are witnessing a leap in scale of a full order of magnitude.

Rather than retiring a system that still has significant scientific value, EPFL has donated Helvetios to help strengthen high-performance computing capacity at the Universidad Nacional de Córdoba (UNC) in Argentina. Once operational, it will triple UNC’s computing capacity, providing researchers with substantially expanded storage, networking, and processing capabilities. Access to more modern computing infrastructure is essential to sustaining scientific research. In many Latin American institutions, major investments in supercomputing are often followed by years without significant upgrades, gradually limiting the scale of research that scientists can undertake. “Without new hardware, technical staff lose interest, scientists stop running computations, and a downward spiral begins. I’ve seen this story play out dozens of times across Latin America: one major investment followed by a decade of declining relative computing performance due to Moore’s Law”, says Nicolas Wolovick, professor and director of high-performance computing at UNC. “This cluster represents a massive surge in computing power for our institution, in scale, storage, and networking capabilities. After just one day, the mindset of our technical team has shifted completely: they are witnessing a leap in scale of a full order of magnitude.”

A Sustainable Approach to Infrastructure Management

For EPFL, the donation also reflects a broader strategy to extend the lifetime of research infrastructure whenever possible. The initiative was developed in collaboration with EcoCloud , EPFL’s center for sustainable digital technologies, which highlighted the environmental benefits of reusing the cluster instead of recycling it. “Antonio Javier Russo, team lead of the computing platforms at EPFL, also played an essential role by establishing the necessary links with partners in Argentina”, says David Atienza, professor and associate vice-president for centers and platforms. He sees the project as a model for future infrastructure management. “We found a win-win strategy, and we hope to replicate this approach many times in the future. We will make sure the computing infrastructure of our new Integrated Computing Platform (ICP), due to launch in January 2027, can be used for as long as possible to minimize the environmental impact of producing new computing systems.”

Transporting the supercomputer from Switzerland to Argentina was a project in itself. “The shipment was made possible through financial support from the Rectorate of the Universidad Nacional de Córdoba., says Dante Paz, astronomy researcher who oversaw the transportation process. Nearly 4 metric tonnes distributed in 8 boxes of equipment—including 288 compute nodes, each equipped with two Intel Xeon Gold 6140 processors and 192 GB of RAM—were shipped by sea. “All the equipment arrived in excellent condition. The packaging worked perfectly, and there were no issues during transport.”

The Helvetios cluster is an ⁠Intel Xeon-based supercomputer. The system delivers a peak performance of 763 TFLOPs across a total of 10,368 processing cores and 54 TiB of RAM. For high-speed temporary storage, it features a 186 TiB all-flash scratch file system capable of reaching a maximum I

O bandwidth of 40 GiB

The core computational power comes from 288 compute nodes. Each individual compute node is equipped with two 18-core Intel Skylake processors operating at 2.3 GHz, 192 GB of DDR3 RAM, and a local 800 GB NVMe drive.

Article on UNC’S website: https:

Source: EPFL centers and platforms

All EPFL centers and platforms news

For years, the Helvetios Supercomputer powered research across EPFL, providing the computing resources needed for data-intensive applications ranging from drug discovery and engineering to cosmological simulations. Now superseded by EPFL’s distributed high-performance computing infrastructure, the supercomputer is embarking on a second career in Latin America. This cluster represents a massive surge in computing power for our institution, in scale, storage, and networking capabilities. After just one day, the mindset of our technical team has shifted completely: they are witnessing a leap in scale of a full order of magnitude. Nicolas Wolovick, professor and director of high-performance computing at Universidad Nacional de Córdoba Rather than retiring a system that still has significant scientific value, EPFL has donated Helvetios to help strengthen high-performance computing capacity at the Universidad Nacional de Córdoba (UNC) in Argentina. Once operational, it will triple UNC’s computing capacity, providing researchers with substantially expanded storage, networking, and processing capabilities. Access to more modern computing infrastructure is essential to sustaining scientific research. In many Latin American institutions, major investments in supercomputing are often followed by years without significant upgrades, gradually limiting the scale of research that scientists can undertake. “Without new hardware, technical staff lose interest, scientists stop running computations, and a downward spiral begins. I’ve seen this story play out dozens of times across Latin America: one major investment followed by a decade of declining relative computing performance due to Moore’s Law”, says Nicolas Wolovick, professor and director of high-performance computing at UNC. “This cluster represents a massive surge in computing power for our institution, in scale, storage, and networking capabilities. After just one day, the mindset of our technical team has shifted completely: they are witnessing a leap in scale of a full order of magnitude.” A Sustainable Approach to Infrastructure Management For EPFL, the donation also reflects a broader strategy to extend the lifetime of research infrastructure whenever possible. The initiative was developed in collaboration with EcoCloud , EPFL’s center for sustainable digital technologies, which highlighted the environmental benefits of reusing the cluster instead of recycling it. “Antonio Javier Russo, team lead of the computing platforms at EPFL, also played an essential role by establishing the necessary links with partners in Argentina”, says David Atienza, professor and associate vice-president for centers and platforms. He sees the project as a model for future infrastructure management. “We found a win-win strategy, and we hope to replicate this approach many times in the future. We will make sure the computing infrastructure of our new Integrated Computing Platform (ICP), due to launch in January 2027, can be used for as long as possible to minimize the environmental impact of producing new computing systems.” © 2026 Universidad Nacional de Córdoba A Smooth Journey Transporting the supercomputer from Switzerland to Argentina was a project in itself. “The shipment was made possible through financial support from the Rectorate of the Universidad Nacional de Córdoba., says Dante Paz, astronomy researcher who oversaw the transportation process. Nearly 4 metric tonnes distributed in 8 boxes of equipment—including 288 compute nodes, each equipped with two Intel Xeon Gold 6140 processors and 192 GB of RAM—were shipped by sea. “All the equipment arrived in excellent condition. The packaging worked perfectly, and there were no issues during transport.” Helvetios at a glance The Helvetios cluster is an ⁠Intel Xeon-based supercomputer. The system delivers a peak performance of 763 TFLOPs across a total of 10,368 processing cores and 54 TiB of RAM. For high-speed temporary storage, it features a 186 TiB all-flash scratch file system capable of reaching a maximum I

O bandwidth of 40 GiB

s. The core computational power comes from 288 compute nodes. Each individual compute node is equipped with two 18-core Intel Skylake processors operating at 2.3 GHz, 192 GB of DDR3 RAM, and a local 800 GB NVMe drive.