CytoNet presented at this year’s ISC conference

The Jülich exascale computer JUPITER was a major topic at this year’s ISC High Performance conference in Hamburg. Ian Buck, Vice President and General Manager of Hyperscale and HPC at NVIDIA, presented research projects carried out on the JUPITER system in an executive briefing—projects that NVIDIA also highlights in a dedicated blog post. Among them is “CytoNet: a foundation model for microarchitecture analysis.” The project is led by Prof. Katrin Amunts and Dr. Christian Schiffer from INM-1.

CytoNet uses imaging data from different cellular scales of the brain to model relationships between individual cell structures and higher-level patterns of brain organization and function. The foundation model was trained on JUPITER in less than five days, using 6.5 petabytes of data from 21 postmortem brains on 4,096 NVIDIA Grace Hopper Superchips.

“For the first time, we’re not just using AI to analyze the brain — we’re building an agent that can think through the experiment itself”, Katrin Amunts, Director of INM-1 and of the C. & O. Vogt Institute for Brain Research at HHU, is quoted as saying in the blog post. This is made possible by JUPITER’s immense computing power and is set to transform the future of neuroscience.

NVIDIA Blog
About CytoNet
Press release of Forschungszentrum Jülich (24 June, 2026): "JUPITER: Exascale Supercomputer with Outstanding Efficiency"