The global change of the entire Earth system makes the understanding of the underlying processes a topic of eminent social-economic importance. Key to this understanding is the modelling of the Earth system, namely the simulation of physical, chemical, biological und geological processes and in particular their interactions. This shall pave the way to the development of prevention and adaptation strategies in the course of global change in order to achieve the sustainable development goals. It is the goal of Earth system modelling to simulate the variability of the Earth system and the underlying processes on different spatial and temporal scales with high precision and to analyse the results and uncertainties on time scales relevant for the management of natural resources and for the regulation by humans and societies.
These simulations require enormous amounts of compute power, especially when interactions between subsystems (e.g. atmosphere, ocean, land surface, cryosphere, solid Earth …) have to be considered, high-precision results are necessary or long-running and large-scale frontier simulations need to be performed over extended time periods at the climate or geologic time scale. Big general-purpose clusters of the Tier-2 performance class with conservative multi-core processor technology, as they are available e.g. at the German Climate Computing Center (DKRZ), are welcomed systems when conducting Earth system simulations. However, these capacity systems cannot perform all the aforementioned simulations in reasonable times. Therefore, Tier-0/1 high-performance computing (HPC) resources with additional advanced many-core technology (e.g. accelerators) must be available to the Earth system science community, in order to provide new impulses to Earth system modelling and consider more complex processes of the Earth system with an increased precision level following the DKRZ expert panel recommendation in January 2016.
The Jülich Supercomputing Centre (JSC) at Forschungszentrum Jülich is the only Helmholtz Tier-0/1 supercomputing centre. Together with its partner institutions Leibniz Supercomputing Centre in Garching and High Performance Computing Center in Stuttgart JSC forms the Gauss Centre for Supercomputing (GCS) which provides and operates the German Tier-0/1 infrastructure. On the European level, JSC plays an important role in the Partnership for Advanced Computing in Europe (PRACE, since 2007), in the EU flagship Human Brain Project (HBP, since 2012), and in four of eight EU-funded Centres of Excellence for computing applications (CoE, since 2015). JSC possesses unique experiences in technology development, maintains long-lasting partnerships with well-known hardware companies, like IBM, Intel and NVIDIA and has access to latest technology through prototype systems. In order to support users of its supercomputing infrastructure in a most efficient way, JSC has introduced Simulation Laboratories (SimLabs) – support and research units for specific scientific communities. Two of these SimLabs are dedicated to Earth system science: the SimLab Climate Science and the SimLab Terrestrial Systems. In summary, this qualifies JSC not only as the one and only Tier-0/1 infrastructure provider but also as a topically well-attached scientific partner of the Earth system science in Helmholtz.
In 2017 Helmholtz decided to fund an additional, dedicated compute partition on the next generation Tier-0/1 system in Jülich (a cluster system being installed mid-2018 and substantially extended by accelerator hardware in 2020; funded by GCS through BMBF and MIWF). Key idea of this engagement is to provide dedicated next generation Tier-0/1 capability to Earth system science for next generation scientific software implementation, testing, performance analysis and tuning on the one hand and for production runs and preparation of ensuing frontier simulations on the other hand. This creates enormous flexibility: Helmholtz Earth system scientists can utilize the dedicated partition under their own governance, while at the same time they can make use of the full capability of the Tier-0/1 system at JSC, e.g. to perform frontier simulations. This concept offers strong synergies with respect to code optimisation and project distribution within a fully equipped HPC eco system.