Digital information technology is developing in a breath-taking fashion, leading to an enormous increase in energy consumption, which is already around 7% of the global electrical energy production. Further scaling of existing information technology is reaching fundamental physical and economical barriers, strongly limiting further increases of its performance.
Neuromorphic computing is one of the potential candidates to replace part of the existing high performance computing architecture, leading to sustainable development of much more energy-efficient scientific computing (Green HPC) with impact to diverse research fields. With a multidisciplinary approach, this project will quantify the actual and potential energy reduction of scientific computing with current and near-commercial neuromorphic hardware. Moreover, the project will benchmark the potential of new neuromorphic hardware platforms for scientific computing.