SpiNNcloud will provide the University of Leipzig with the world's largest neuromorphic supercomputer. It will include 656,640 cores and will be used for modeling proteins and other tasks in the field of personalized medicine.
The system is based on principles of the human brain. Neuromorphic architecture allows for efficient data processing by mimicking neural networks. The system is based on 4,320 SpiNNaker2 (Spiking Neural Network Architecture) ARM chips, each containing 152 cores. This number allows for modeling up to 10.5 billion neurons.
The server architecture involves 48 chips per board, providing high computing density and low power consumption. It utilizes many low-power CPUs instead of powerful but power-hungry GPUs.
According to the developers, the new platform consumes 18 times less energy than modern GPU solutions of comparable computing power. The project, estimated to cost several million euros, represents a significant step forward in the development of sustainable technologies for the future of medicine.
