ESCAPE will develop world-class, extreme-scale computing capabilities for European operational numerical weather prediction (NWP) and future climate models.
To achieve this, ESCAPE has set five top-level objectives:
- Define fundamental algorithm building blocks (hereafter called “Weather & Climate Dwarfs”) to foster trans-disciplinary research and innovation and to co-design, advance, benchmark and efficiently run the next generation of NWP and climate models on energy-efficient, heterogeneous HPC architectures
Achieved by: Breaking down NWP legacy codes into Weather & Climate Dwarfs, in analogy to the Berkeley Dwarfs (Asanovic et al., 2006), and providing mini-applications (Heroux et al., 2009) for all performance relevant model parts.
- Diagnose and classify Weather & Climate Dwarfs on different HPC architectures.
Achieved by: Implementing and analysing prototype dwarfs using the following categories: data flow and data locality, parallel communication pattern, energy-aware performance metrics, floating point rates, and limits of accuracy and resilience (e.g. to handle lower precision and/or specific hardware faults) required from software and hardware for the specific task of NWP simulations, thus guiding both future hardware and compiler development.
- Combine frontier research on algorithm development and extreme-scale, high-performance computing applications with novel hardware technology, to create a flexible and sustainable weather and climate prediction system.
Achieved by: Optimally mapping key NWP processes to energy-efficient,specialized compute units and novel accelerator technologies, addressing performance portability, and by establishing in NWP novel data structures, mathematical algorithms and numerical methods encapsulated in new dwarfs.
Foster the future design of Earth-system models and commercialisation of weather-dependent innovative products and services in Europe through enabling open-source technology.
Achieved by: Providing an open-source data structure framework to fundamentally reform the way Earth-system modelling is performed, disseminating the concepts and providing training, and addressing the reintegration of optimized Weather & Climate Dwarfs into legacy codes.
Pairing world-leading NWP with innovative HPC solutions and thus “taking great ideas from the lab to the market”, fostering economic growth, EU business sector competitiveness and creation of new jobs.
Achieved by: Directly involving established high-performance computing (HPC) and a European start-up Small and Medium-sized Enterprise (SME), the business sector, and supporting them in tailoring future market solutions while minimizing energy requirements, and at the same time, enabling very high-resolution NWP models as tools for planning and decision processes (e.g. power production prediction) for wind energy, wave and tidal motions, and solar power.