It is becoming increasingly apparent that the future energy ecosystem will rely heavily on digitization to drive essential innovations in production and storage technologies, mitigate power source variability and manage its distribution via a complex hierarchy of micro- and macro-networks.
The EoCoE consortium is composed of world-leading research teams from four strategically important low-carbon energy domains in meteorology, materials, hydrology and fusion, linked together through a multi-disciplinary platform of high performance computing (HPC) and numerical mathematics to create a network of experts in computational energy science.
The EoCoE consortium developed a comprehensive, structured support pathway for enhancing the HPC capability of energy-oriented numerical models, from simple entry-level parallelism to fully-fledged exascale readiness. At the top end of this scale, promising applications from each energy domain have been selected to form the basis of 5 new Energy Science Challenges in the present successor project EoCoE-II that will be supported by 4 Technical Challenges as represented on the following figure.
The Science Challenges cover the following areas:
The Technical Challenges cover:
- Scalable Solvers
- Programming Models
- IO & Data Flow
- Ensemble Runs
The project organization can be seen as a Matrix structure where the 4 Technical Challenges (shown in the rows) are transversally interlinked to the 5 Scientific Challenges so that the HPC experts will work in a very integrated manner with the domain scientists.
In order to define the work packages (WP), we chose to group tasks by technical challenges rather than by scientific discipline as shown in the following figure. WP2 to WP5 represents the Technical Challenges. The main advantage of this choice is a common HPC and mathematical background of the work package leaders, which eases a lot the communication between WPs and enhances the project’s receptiveness to exascale hardware developments. Nevertheless, special care has been taken to ensure that the domain scientists retain leading roles in the project in order to keep it user-driven. The Exascale Co-design Group (ECG) [link toward the ECG section], a key component of the management structure in WP7, ensures that technological and design decisions are made by the code owners in close consultation with HPC experts. WP1 is responsible for the scientific challenges and applications. WP6 is dedicated to the dissemination and the networking.
In each Scientific Challenge, we have decided to focus refactoring and optimization on a limited number of flagship applications (see Scientific Challenge pages for details).
Flagship applications have different needs and will therefore not be equally represented in the Work Packages. The work to be done in each of them is schematically represented in the following figure.
A detailed description of each Work Package and their contribution to the different codes is given in the different tabs.
In addition to the flagship applications, so-called satellite codes will be used to provide some data, to be coupled with the flagship codes or to do some result comparison. Satellite codes will also be improved during the project or will benefit from WP expertise. External codes will be used but not modified within the project.