Caught between the continuous technological advances towards exascale computation and the global efforts for greener IT, power consumption of modern High Performance Computing (HPC) systems has become the major challenge on the road to exascale computing. The prospect of exascale machines finally enabling the quantification of uncertainties related to multi-scale modelling and simulation inevitably leads to the imperative need to reduce the prohibitive energy requirements of HPC systems.
However, as the current efforts towards an efficient reduction of energy consumption mainly focus on re-engineering facilities and hardware, the great potential of transforming and completely re-designing algorithms implemented in the applications on HPC systems remains so far unexplored.
In the Exa2Green project, an interdisciplinary research team of HPC experts, computer scientists, mathematicians, physicists and engineers takes up the challenge to develop a radically new energy-aware computing paradigm and programming methodology for exascale computing. As a proof of concept, the online coupled model system COSMO-ART, based on the operational weather forecast model of the COSMO Consortium (www.cosmo-model.org), is being modified to incorporate energy-aware numerics. COSMO-ART was developed at KIT and allows the treatment of primary and secondary aerosols and their impact on radiation and clouds.
"The concept of energy-aware numerics is a multi-level approach in which all layers from kernels, numerical building blocks to application level are designed by means of an optimised trade-off between energy consumption, performance and accuracy."