The LUMI Supercomputer specs, 3 World-Changing Applications

LUMI Supercomputer

EuroHPC JU, an EU program to build and deliver exascale computing platforms, funds the LUMI supercomputer in Kajaani, Finland. The money-giving LUMI consortium includes Finland, Belgium, Czech Republic, Denmark, Estonia, Iceland, the Netherlands, Norway, Poland, Sweden, and Switzerland.

In November 2024, the Top500 list named LUMI the fastest supercomputer in Europe and the sixth fastest in worldwide. At its peak, the LUMI Supercomputer can handle 380 petaflops, or 1.5 million high-end laptops. The AMD EPYC CPUs and AMD Instinct MI250X GPUs form its foundation on the HPE Cray EX platform. The Green500 list ranks LUMI as the 25th most energy-efficient supercomputer in the world. Operating entirely on hydropower, the facility’s waste heat is recovered and used to heat roughly 100 homes in Kajaani.

Here are three ways teams are leveraging LUMI’s enormous processing capacity to impact the world, though it can support a lengthy list of initiatives.

Addressing climate change

One of the greatest challenges facing humanity is determining how to slow down the trajectory of global warming and address its effects. Destination Earth is among the most fascinating climate-related projects that utilize LUMI. Five kilometers of the planet are precisely modeled by this digital twin project. This initiative produces approximately one petabyte of data every day, which is too large to handle without a supercomputer.

Once every seven to ten years, climate experts could revise their projections of the effects of climate change. The ability to release updated projections once or more a year with Destination Earth has significantly improved prediction accuracy, enabling humanity to get ready.

Scientists can also examine what might occur under a number of different scenarios with the model. Destination Earth, for instance, enables government organizations to model the consequences of proposed weather-related policies to determine how successful they will be in minimizing or lowering the effects of global warming.

Forecasting natural disasters

In conjunction with Destination Earth, LUMI can predict droughts, floods, storms, and cyclones. Climate change killed at least 3,700 people in 2024. Scientists expect that even as the number of disasters rises, improved analytical powers will help them lower that number.

The LUMI supercomputer is also a part of the European High-Performance Computing Centre of Excellence for Exascale on Solid Earth (ChEESE CoE) and the Destination Earth project. Accurately forecasting glacial outburst floods is one project. One of the countries in the LUMI supercomputer cooperation, Iceland, is at high danger from glacier outburst flooding. In regions that are physically challenging to reach, a number of variables, including temperature, weather, glacier melt, and minor seismic events, interact in intricate ways, making these events infamously unpredictable.

One of the LUMI consortium countries, Iceland, is at serious risk from these floods. Using Elmer/Ice software on the LUMI system, scientists from multiple organizations are simulating these abrupt glacier outburst floods that threaten Icelandic villages in an effort to provide early warnings that could save lives.

The ChEESE CoE and LUMI are also working on additional projects that analyze hydrology, earthquakes, fires, volcanic ash, and more. The LUMI system has cycles specifically designated for urgent computing, and LUMI is working on projects that use artificial intelligence to weather forecasting.

Treating disease

LUMI is a vital tool for life sciences research since it offers information on how to diagnose and treat illnesses. ComPatAI was one of its initial initiatives; it analyzes tissue samples using deep learning methods. “We have primarily worked on breast and prostate cancer,” says Pekka Ruusuvuori, an associate professor at the University of Turku’s Institute of Biomedicine. Since these are the most prevalent malignancies in both men and women, there is a wealth of information available on them. However, goal is to develop a highly versatile model that may be further enhanced for novel and diverse applications.

The leading provider of healthcare lab services in Finland, Fimlab, provides digitized slides to the initiative. By project completion, ComPatAI anticipates 2.5 million photos. Because of this amount of data, they are able to train AI models to identify cancer that is invisible to the human eye.

Another effort uses LUMI to assist in the creation of atomic-scale cell membrane models. Cell membrane alterations can be used by doctors to diagnose and treat type 2 diabetes, Parkinson’s disease, and Alzheimer’s disease. Because atomic-scale physical tests are difficult to conduct to learn about healthy and diseased cells, scientists are using simulations to improve their understanding, which could alter medicine.

LUMI Supercomputer Specs