Launched in 2023 for a duration of 6 years, The NumPEx PEPR aims to contribute to the design and development of numerical methods and software components that will equip future European Exascale and post-Exascale machines. NumPEx also aims to support scientific and industrial applications in fully exploiting their potentials.

Exa-MA aims to revolutionize methods and algorithms for exascale scaling: discretization, resolution, learning and order reduction, inverse problem, optimization and uncertainties. We are contributing to the software stack of future European computers.

Taking into account multiple and coupled physics is at the heart of many application requirements in fields as varied as, but not limited to, aeronautics, defense and biology. This is also strong area of expertise for CEA Energy Division, with multiple domains including fluid-structure interaction, neutronics coupled with thermal-hydraulics a/o thermal-mechanics or severe accident modeling. The emergence of exascale architectures opens the way to promising new levels of high-fidelity simulations, but also significantly increases the complexity of many software applications in terms of total or partial rewriting. Therefore, it specifically encourages partitioned coupling to limit development work. The idea is to search for each physics of interest in a necessarily reduced number of highly optimized software components, rather than making specific, possibly redundant developments in standalone applications.
Once the coupled multiphysics problem has been written with the expected levels of accuracy and stability, the proposed work concentrates on the coupling algorithms between applications assumed to be themselves exascale-compatible, to be solved efficiently at exascale. It is also worth noting that, in general, the couplings under consideration can present a high level of complexity, involving numerous physics with different level of feedback between them and various communications from exchanges of boundary conditions to overlapping domains. The current post-doctoral internship, to be carried out in the framework of the Exa-MA collaborative project, is in particular dedicated to the identification and dynamic tuning of the relevant numerical parameters arising from the coupling algorithms and impacting the computational efficiency of the global simulation. Considered problems are in the general case time-evolving problems, with a significant number of time iterations, allowing using the first iterations to gather data and conduct the tuning.
Regarding software, the research is intended to be conducted within the ICoCo/C3PO low-intrusive open-source coupling framework, with data exchange through the MEDCoupling library. To benefit from the synergies made possible by the Exa-MA context, it is foreseen to mutualize autotuning techniques with related work addressing adaptive precision and validation in the Promise software environment.

Two main topics can be identified in terms of computational efficiency specific to coupling frameworks: load balancing and scheduling on the one hand, and coupling algorithms and data transfers on the other hand. The proposed internship is dedicated to the second topic and aims at designing a strategy to automatically identify, rank and optimize the internal parameters impacting the cost of the coupling computational tasks. Such an autotuning approach is mandatory to provide the necessary versatility in the coupling tools to adjust to any multiphysics configuration with minimal a priori knowledge of its characteristics. Expected parameters to be tuned this way can be, without prejudging the findings of the proposed research, convergence criteria in the loops between physics, computing orders between physics, as well as possible multi-level convergence between groups of physics.
Practically, the work is intended to be based, in a first phase, on model apps mimicking the behaviour of several configurations of multiphysics coupling, in terms of computational cost of each physics and feedback between them, to design and test autotuning strategies for the automated discovery of a known optimal execution path. The second phase will consist in practical benchmarks between actual multiphysics tools on well chosen problems of interest, to quantify the performances and improve the strategies designed before. Fluid & structure (thermal-) mechanics and neutronics are primarily targeted physics for the tests, without excluding any additional coupling proving worthy of studying during the internship.

Candidates must have a PhD in Computer Science, Applied Mathematics or other relevant fields.

Good programming skills are required.