Maison de la Simulation is recruiting a Research Engineer specialised in high-performance computing. Please visite CNRS recruit page for more information.
Laboratory: Maison de la Simulation (UAR3441, CEA Saclay, 91191 Gif-sur-Yvette Cedex)
CNRS institute: INS2I
Recruitment process opening date: June 2022
Official offer webpage: Concours N° 49
Maison de la Simulation (www.mdls.fr) is a joint-laboratory (CNRS, CEA, Université Paris-Saclay, UVSQ) whose mission is to support scientific communities to efficiently exploit current and future supercomputers. Future supercomputers, in the context of the European Exascale project, will be equipped with innovative technologies for high-performance computing. New processors based on the RISC and ARM CPU architectures will emerge. Super-computing facilities will be equipped with a record number of compute nodes and heterogeneous nodes where the computing power will be concentrated on accelerators, GPUs, or FPGAs for example. You will join the multidisciplinary team of Maison de la Simulation, one of whose strengths is to bring together many profiles around high-performance computing: academics, researchers, engineers, from the world of computer science, physics, or applied mathematics.
As part of the Maison de la Simulation team, one of your main missions will be to bring your expertise in software engineering and high-performance computing to the Smilei simulation code (https://smileipic.github.io/Smilei/index.html). This code is intended for the numerical study of plasmas in extreme conditions, such as those encountered on so-called extreme light laser installations like the Apollon laser, or in certain astrophysical environments (solar wind, super-novae, pulsar, etc.) which require the performance of large-scale numerical simulations on the most advanced machines (GENCI and PRACE). It is a massively parallel open-source code, mostly written in C++ and used by many scientific teams around the world. Your short- and medium-term role will be to lead the HPC developments needed for the arrival of the next European exascale supercomputers (ARM optimisation, GPU porting). Your long-term role will be to become the main code architect. You will be responsible for the evolution and modernisation of its overall structure.
You will carry out your mission in close collaboration with the scientific development teams present on the Saclay plateau (LLR and LULI laboratories at Ecole Polytechnique). You will contribute to the writing of scientific projects related to high performance computing and in this context, you will participate in the supervision of trainees, PhD students and engineers.
Your work will be carried out in connection with the other scientific and technical activities of Maison de la Simulation. You will also carry out a technological watch activity and take part in the study of new emerging technologies for high-performance computing (new processors, new accelerators, new programming models, future quantum coprocessors, models based on automated learning, processing of massive data). Your expertise will reinforce the pool of competencies of the laboratory.
According to your interests, you will lead training, animation, communication, and scientific popularisation activities of the laboratory activities.
- Lead application developments on the Maison de la Simulation’s flagship research themes to make full use of the major computing infrastructures at national or European level,
- Provide high-level expertise in the field of high-performance computing to researchers and advise them on the use of large-scale computing resources,
- Participate in European and national projects in areas related to HPC applications, including project management and staff supervision,
- Carry out training activities in the laboratory and ensure the transfer of knowledge and expertise to scientists (design and organisation of courses and training),
- Ensure a technological watch on emerging developments of the field (new hardware architectures, new programming models, etc.).
- Significant expertise in the development of parallel applications in distributed or shared memory,
- Optimisation of application software on various processor architectures (CPU, GPU, FPGA),
- Operational knowledge of advanced programming techniques and languages (C++ in particular),
- Knowledge of GPU accelerator programming techniques (e.g., OpenACC or CUDA) or new programming models (Kokkos, SYCL, oneAPI) will be a plus,
- English language: fluency in written and spoken scientific English (B1 to B2, Common European Framework of Reference for Languages).