The LUMI Supercomputer specs, 3 World-Changing Applications
LUMI Supercomputer
EuroHPC JU funds the Kajaani, Finland-based LUMI supercomputer. Finland, Belgium, Czech Republic, Denmark, Estonia, Iceland, the Netherlands, Norway, Poland, Sweden, and Switzerland support LUMI.
Top500 named LUMI Europe's fastest supercomputer and sixth overall in November 2024. The peak LUMI Supercomputer can handle 380 petaflops, or 1.5 million high-end laptops. HPE Cray EX is built on AMD EPYC CPUs and AMD Instinct MI250X GPUs. LUMI is the 25th most energy-efficient supercomputer on Green500. The hydropower-powered plant recovers waste heat to heat 100 Kajaani dwellings.
Though LUMI may assist many efforts, here are three ways teams are using its massive processing capability to change the world.
Addressing climate change
How to halt global warming and mitigate its impacts is one of humanity's biggest issues. Destination Earth is one of the most intriguing LUMI climate initiatives. The digital twin project accurately models five km of the earth. A supercomputer is needed to manage this initiative's daily petabyte of data.
Every seven to ten years, climate experts might update their climate change estimates. Destination Earth's annual estimates have increased prediction accuracy, allowing mankind to prepare.
The model lets scientists test alternative situations. Destination Earth allows government agencies to simulate weather-related programs to measure their effectiveness in reducing global warming.
Forecasting natural disasters
LUMI and Destination Earth anticipate droughts, floods, storms, and cyclones. At least 3,700 died in 2024 from climate change. Scientists predict better analytical skills to reduce disasters as they increase.
The Destination Earth project and European High-Performance Computing Centre of Excellence for Exascale on Solid Earth (ChEESE CoE) use the LUMI supercomputer. Creating accurate glacier outburst flood forecasts is one endeavour. Glacier outburst floods threatens Iceland, a LUMI supercomputer cooperation member. In difficult-to-reach places, temperature, weather, glacier melt, and tiny seismic occurrences combine in complex ways, making these events notoriously unpredictable.
These floods threaten Iceland, a LUMI consortium member. Using Elmer/Ice software on the LUMI system, scientists from several organisations are modelling sudden glacier outburst floods that endanger Icelandic settlements to offer lifesaving early warnings.
Other ChEESE CoE and LUMI initiatives analyse hydrology, earthquakes, fires, volcanic ash, and more. LUMI has urgent computer cycles and is working on weather forecasting projects using AI.
Treating illness
LUMI provides disease diagnosis and treatment information, making it essential for life sciences research. One of its early projects, ComPatAI, analyses tissue samples using deep learning. “We have primarily worked on breast and prostate cancer,” says University of Turku Institute of Biomedicine associate professor Pekka Ruusuvuori. These are the most common cancers in men and women, therefore there's much of information about them. The objective is to provide a flexible model that can be expanded for new and different uses.
Finnish healthcare lab leader Fimlab offers digitised slides to the program. By project completion, ComPatAI expects 2.5 million images. With this much data, they can train AI algorithms to detect unseen cancer.
Another project employs LUMI to model atomic-scale cell membranes. Cell membrane modifications can help doctors identify and treat type 2 diabetes, Parkinson's, and Alzheimer's. Scientists employ simulations to study healthy and sick cells since atomic-scale physical testing are impractical. This might change medicine.