​High Performance Computing (HPC) is the basis for numerical simulation and for big data processing, which are key enablers for excellence in science and industrial competitiveness and economic growth.

France and Japan both have national projects for next generation supercomputers (2020- 2025), and foster co-design with industrial partners and research organizations. There is also a strong convergence of viewpoints on technological and methodological options: ARM processor technology, open software ecosystems interoperability between architectures, co-design and partnership with industrial technology suppliers, and development of user ecosystems.
After one year of intensive collaboration in HPC, the French and Japanese teams, from different CEA divisions (France) and from RIKEN R-CCS, are now foreseeing the active continuation of this productive work and jointly looking forward to perspectives for 2020 and beyond. The challenges to be tackled for this purpose encompass energy-efficiency of the computing system; their ease-of-use and overall productivity – how they can be programmed to enable important applications for science and society; the cost effectiveness of the solutions.
The CEA and RIKEN HPC collaboration covers the development of open source software components to address these challenges, from middleware and operating system level to applications, as well as training efforts and reflections on benchmarking. The developments are organized to benefit both mainstream and emerging HPC architectures (e.g. x86 as well as ARM-based).
 
The CEA and RIKEN HPC experts met in Paris early March 2018 for a plenary meeting and progress review of their collaboration. The following topics of the collaboration were discussed during the two days:

• programming environments and languages, runtime environments, and energy-aware job schedulers technical topics;
• the performance and effectiveness indicators and metrics– with the central concern of designing useful and productive computers with the best possible efficiency;
• human resources, training and skill generation
• lastly, the first applications proposed in the collaboration, dealing with quantum chemistry and condensed matter physics, and the seismic resistance of nuclear facilities.

Detailed multi-annual tracks of action had been defined for each topic, and the work plan for Year 1 has been fulfilled. In addition to extensive mutual information and setting up different threads of action, some salient achievements can already be highlighted, such as extensions and coupling of CEA and RIKEN programming environments (resp. so-called MPC and XMP).