At ISC, JUPITER Shows What Exascale Science Looks Like
AI Summary: JUPITER, Europe’s first exascale supercomputer, has enabled significant advancements in four key scientific projects, showcasing the capabilities of exascale computing. Notably, the Jülich Brain Atlas project has developed CytoNet, a foundation model for analyzing brain microarchitecture, utilizing 6.5 petabytes of data from 21 post-mortem brains and trained in under five days on JUPITER. Additionally, the ICON model, which won the Gordon Bell Prize, simulates the Earth’s climate at a 1-kilometer resolution, allowing for comprehensive modeling of interconnected ecosystems and achieving a world record in global climate simulation. These projects demonstrate the potential of exascale computing to tackle previously intractable scientific challenges.