Suivez

la liste

We are seeking to fill a 24-month post-doctoral position to work at the Saint-Venant Laboratory for Hydraulics (LHSV) in Chatou (Paris area, France) to contribute to the performance optimization of the openTELEMAC system on modern CPU and GPU architectures, with the aim of accelerating high-fidelity free-surface flow simulations for environmental and coastal applications. Details on the research project, hosting laboratory and conditions to apply are given below.

General context of the research project:

In its 35 years of existence and development, the openTELEMAC software suite1 has become a well-established tool for assessing environmental conditions in seas, estuaries, coastlines, lakes and rivers. Its current parallelization strategy relies on MPI, targeting CPU architectures exclusively. Over the past decade, GPU processing power has increased considerably and notably at a much faster rate than that of CPUs. Integrating GPU computing into advanced simulation software such as openTELEMAC provides a compelling opportunity to apply and further advance code optimization expertise. This postdoctoral position will involve CPU-side performance analysis, examination of data structures and access patterns, code optimization, and, where significant performance gains are anticipated, the porting of relevant code components on GPUs.

Subject and objectives of the post-doctoral research:

What you can expect:

  1. Experience in software development in an industrial context and under industrial constraints.
  2. Experience with performance analysis tools on different HPC architectures.
  3. Analysis of data structures and access patterns in memory-bound settings.
  4. Implementation and performance analysis of GPU prototypes for performance critical functionalities.
  5. Experience of multi-language, multi-paradigm parallel programming (Fortran + X, MPI, OpenMP).
  6. Integration into a multi-disciplinary development team, working on a code relevant for environmental studies.

What we expect of you:

  1. Curiosity.
  2. Creativity to identify possibilities for optimization in an industrial (constraint) setting.
  3. Prior experience in scientific computing in general and numerical PDEs in particular.
  4. Solid programming experience in a compiled language (C++, Fortran, ...) on distributed memory architectures.
  5. Willingness to work with legacy code in Fortran.

References:

  • C. Moulinec, C. Denis, C.-T. Pham, D. Rougé, J.-M. Hervouet, E. Razafindrakoto, R. W. Barber, D. R. Emerson, and X.-J. Gu, "TELEMAC: An efficient hydrodynamics suite for massively parallel architectures", Computers & Fluids, vol. 51, no. 1, pp. 30–34, 2011. https://doi.org/10.1016/j.compfluid.2011.07.003
  • J. Grasset, S. M. Longshaw, C. Moulinec, D. R. Emerson, Y. Audouin, and P. Tassi, “Porting TELEMAC-MASCARET to OpenPOWER and experimenting GPU offloading to accelerate the TOMAWAC module”, in Proc. TELEMAC User Conference 2019, Toulouse, France, 2019. https://doi.org/10.5281/zenodo.3611548
  • H. Anzt, M. Baboulin, J. Dongarra, Y. Fournier, F. Hulsemann, A. Khabou, and Y. Wang, “Accelerating the Conjugate Gradient Algorithm with GPUs in CFD Simulations”, in High Performance Computing for Computational Science – VECPAR 2016, I. Dutra, R. Camacho, J. Barbosa, and O. Marques, Eds., Lecture Notes in Computer Science, vol. 10150. Cham: Springer, 2017, pp. 35–43. https://doi.org/10.1007/978-3-319-61982-8_5
  • E. Malaboeuf, A. Marty, O. Hoenen, and P. Tamain, “Porting a Fortran plasma simulation to Exascale on AMD GPUs using both OpenMP and Kokkos”, in Proc. SC'25 Workshops, ACM, 2025. https://dl.acm.org/doi/10.1145/3731599.3768148
  • M. Bernardini, D. Modesti, F. Salvadore, S. Sathyanarayana, G. Della Rocchetta, and S. Pirozzoli, “OpenMP offload toward the exascale using Intel® GPU Max 1550: evaluation of STREAmS compressible solver”, The Journal of Supercomputing, vol. 80, pp. 18466–18487, 2024. https://doi.org/10.1007/s11227-024-06254-y

About Saint-Venant laboratory (LHSV) and the hosting research team:

The Saint-Venant Laboratory for Hydraulics (LHSV) is a joint research unit between Ecole des Ponts (ENPC), one of the top-level engineering schools in France, and the R&D division of Electricité de France (EDF R&D), Dept. Laboratoire National d’Hydraulique et Environnement (LNHE). The laboratory has a unique position in the field of applied fluid mechanics for free surface flows in inland waters, rivers, estuaries, coastal zones and oceans, at the interface between academy and industry.
The laboratory has access to a variety of experimental facilities as well as two high-end HPC clusters, one of which is ranked in the TOP500 list of the world's most powerful supercomputers, on which openTELEMAC simulations are regularly performed. The laboratory is located on the EDF Lab site in Chatou (France), about 12 km from Paris downtown (6 quai watier, 78400 Chatou). It can be reached by the public transportation network of Paris area (RER line A – station “Rueil-Malmaison”).
More information about our laboratory can be found at: http://www.saint-venant-lab.fr

Who can apply:

By the start of the appointment, successful applicants should hold a Ph.D. in scientific computing or a related field, or equivalent experience, in particular on distributed memory parallel architectures. Prior experience with GPU programming (CUDA, Kokkos, HIP, …) is highly desirable. Familiarity with Fortran is an asset but is not required. The researcher will be employed by Ecole des Ponts (ENPC), with a 24-month contract (nominal weekly working rate of 35 hours). Salary and benefits will be commensurate with qualifications and experience.

How to apply:

Applicants should send a cover letter and CV to Boris BASIC, Cédric GOEURY, Frank HULSEMANN and Bruno LATHUILIERE (see e-mail address below).
For additional information, please contact:

The deadline for application is June 30th, 2026.
The targeted starting date of the post-doctoral contract is September 1st, 2026.