Green Information Technology collaboration: Energy reduction for scientific calculations
Collaborating with SURF Lab is easy
SURF’s Open Innovation Lab is currently working with different parties on several technological innovation projects. Collaboration, trans-disciplinary approach and experimentation are key here. “SURF’s Lab is always open to possible new collaborations”, says Program Lead Sagar Dolas. “We’re making it easy for research groups to participate with us, by moving away from long planning and approval periods. This current project with Radboud University is an example of this way of working. We try to be flexible and start a collaboration if we’re a fit, because we want to engage and inspire. We want to be as inclusive as possible, by engaging both with larger and smaller partners.”
Increased energy consumption
Digital information technology is developing in a breath-taking fashion, leading to an enormous increase in energy consumption, which is expected to reach 7% of the global electrical energy production. Science itself is a key actor in this development: Today, world fastest supercomputer has peak power usage of ~30 MW which is almost equal to electricity consumption of a moderate sized city.
The potential of neuromorphic scientific computing
Neuromorphic computing (emulating neural structure and operation of human brain) is one of the potential candidates to replace parts of the existing scientific calculations. This collaboration focuses on quantifying the energy reduction and benchmark the potential of the hardware platform for scientific research. The project has been awarded 30.000 Euros for 6 months within the Radboud Centre for Green Information Technology, which is currently under establishment.
The role of SURF
SURF Open Innovation Lab will provide in-kind contribution and consultancy to the partners from this consortium to access the cluster for energy measurement specially built for joint experiments of this nature. The team consists of Johan Mentink, Johan Kwisthout, Theo Rasing and Sascha Caron from Radboud University, plus Sagar Dolas (SURF) and Heike Riel (IBM Zurich).