Projects and collaborations

Current projects

Currently, I am involved in two nation-wide projects considered by the government as national priorities for the coming years: PIA/Equipex+ MesoNet and PEPR/Numpex. Equipment of excellence (Equipex) projects are funded through the Investments for the Future Program (PIA), set up in 2011 by the government to support innovative and promising investments. Priority research programs and equipment (PEPR) projects are dedicated to support hot and strategic topics including exascale computing. The PEPR/Numpex nationwide program is part of the French application to EuroHPC’s Call for Expressions of Interest (AMI) (Exascale Project France), to host one of the two European exascale machines planned for Europe by 2024. In addition, I have been involved in three regional projects funded through the CPER tripartite program Region-State-Europe: CornelIA (2021-2027) and Data (2015-2021). Finally, I have an international bilateral project supported by the French ANR national research agency and its counterpart: FNR/Luxembourg. More details on the projects are given in the following.

  • ANR PEPR Numpex/Axis Exa-MA (2022-2027, Grant: Total: 6,5M€). The goal of the high-performance Digital for Exascale (Numpex) nation-wide program, dedicated to both scientific research and industry, is twofold: (1) designing and developing the software building-blocks for the future exascale supercomputers, and (2) preparing the major application areas aimed at fully harnessing the capabilities of these latter. Numpex is composed of 5 axes including Exa-MA, which stands for Exascale computing: Methods and Algorithms and is organized in 7 WPs including Optimize at Exascale (WP5). The overall goal of WP5 consists in the design and implementation of exascale algorithms to efficiently and effectively solve large optimization problems. E-G. Talbi and I are respectively the leader of and a contributor to this work-package.

  • ANR PIA Equipex+ MesoNet (2021-2027, Grant: Total: 14,2M€, For ULille: 1,4M€). The goal of this nation-wide project is to set up a distributed infrastructure dedicated to the coordination of HPC and AI in France. This inclusive and structuring project, supported by GENCI partners (MESRI, CNRS, CEA, CPU, INRIA), aims to integrate at least one mesocenter by region making them regional references and relays. The infrastructure, fully integrated with the European Open Science Cloud (EOSC) initiative, should have a significant impact on the appropriation by researchers of the national and regional public HPC and AI facilities. Coordinated by GENCI, MesoNet gathers 22 partners including the mesocenter located at ULille, for which I am the co-PI. The MesoNet infrastucture is highly important for the research activities of my team and many other research groups. In addition to the funding dedicated to hardware equipment including nation-wide federated supercomputer and storage, funding will be devoted to research engineers, one of them for ULille (4,5 years), and a PhD under my supervision as well.

  • International ANR-FNR France/Luxembourg PRCI UltraBO (March 2023-July 2026, Grant: 207K€ for ULille, PI: N. Melab) in collaboration with University of Luxembourg (Co-PI: G. Danoy). According to Top500 modern supercomputers are increasingly large (millions of cores), heterogeneous (CPU-GPU, . . . ) and less reliable (MTBF < 1h) making their programming more complex. The development of parallel algorithms for these ultra-scale supercomputers is in its infancy especially in combinatorial optimization. Our objective is to investigate the MPI+X and PGAS-based approaches for the exascale-aware design and implementation of hybrid algorithms combining exact methods (e.g. B&B) and metaheuristics (e.g. Evolutionary Algorithms) for solving challenging optimization problems. We will address in a holistic (uncommon) way three roadblocks on the road to exascale: locality-aware ultra-scalability, CPU-GPU heterogeneity and checkpointing-based fault tolerance. Our application challenge is to solve to optimality very hard benchmark instances (e.g. Flow-shop ones unsolved for 25 years). For the validation, various-scale supercomputers will be used, ranging from petascale platforms, to be used for debugging, including Jean Zay (France), ULHPC (Luxembourg), SILECS/Grid’5000 (CPER CornelIA, 2021-2027) and MesoNet (PIA Equipex+, 2021-2027) to exascale supercomputers, to be used for real production, including the the first supercomputer of Top500 (Frontier via our Georgia Tech partner) as well as the two EuroHPC coming ones.

  • CPER CornelIA (2021-2027, Grant: 820K€): this region-wide project aims at strengthening the research and infrastructure necessary for the development of scientific research in responsible and sustainable Artificial Intelligence at the regional (Hauts-de-France) level. As scientific leader at Lille, I am in charge of the management of the the renewal of the hardware equipment of Grid’5000 nation-wide experimental testbed and hiring an engineer for its system & network administration and user support and development.

  • Bilateral industrial-context project with ONERA (2017-2023, Paris): the focus of this project with a major European player in vehicle aerospace is put on the design of aerospace vehicles, an expensive multidisciplinary problem. Two jointly supervised Ph.D students (A. Hebbal and J. Gamot) have been/are involved in this project. The Ph.D thesis of A. Hebbal has been defended in January 2021. The Ph.D thesis of J. Gamot has started in November 2020. The objective of this latter is to deal with the design and implementation of ultra-scale multi-objective highly constrained optimization methods for solving the internal layout problem of future aerospace systems.

  • PyParadisEO (2020-2022, Grant: 27-months engineer funding) is a software development project (ADT) supported by Inria. The objective is to migrate our ParadisEO software framework from C++ to Python (PyParadisEO) to facilitate its use and interoperability with other solvers and tools (e.g. ML). Jan Gmys has been hired in Oct. 2020 to work on the migration task until Dec. 2022.

Previous projects

  • H2020 SYNERGY (Grant agreement ID: 692286) (Feb. 2016-Jan. 2019, Grant: 260K€ / 1M€ for Lille) is a twinning-of-research-institutions European research project with two partners: Jožef Stefan Institute (JSI), Ljubljana, Slovenia and Cologne University of Applied Sciences (CUAS). The overall goal of the project was to improve JSI excellence and unleash its research and innovation potential through training in parallel optimization/computing (by Lille) and surrogate modelling (by CUAS) and networking in the context of multi-objective optimization. I have been the leader of “Training", one of the four WPs of the project. The outcome of this latter includes mainly the joint organization of events such as Synergy summer school at Ljubljana and the PACOS workshop, the involvement in the HPOI conference, the joint editing of a Springer book and LNCS conference (BIOMA) proceedings, etc.

  • ARCUS E2D2 (2014-2018). In collaboration with 3 universities in the Hauts-de-France region (Université de Lille, ULCO and UVHC) and in partnership with Lebanon, Morocco and Palestine, the ARCUS project “E2D2" e2d2.univ-lille1.fr/projets/) aims to federate and develop activities of scientific cooperation in research and training on the themes of Energy, Environment and Sustainable Development. “E2D2" project is coposed of 4 workpackages, including “Scientific Computing Platform" of which I am co-leader.

  • National ACI/ANR projects. In the 2000s, we witnessed the emergence of grid computing technonlogies. Consequently, several national ACI/ANR research programs have been launched. I have been involved, either as a principal investigator or as as the leader of a workpackage on grid-based implementation of parallel optimization methods, in 5 projects related to three of these ACI/ANR programs: 2 projects (DOCK et CHOC) in ANR “Calcul intensif et grilles de calcul", 1 in ACI “Masses de données" (GGM) and 2 in ACI Grid (Grid’5000 and DOC-G). A brief decription of the projects is given in the following.

    • ANR DOCK (2006-2009). This is a collaborative project with two partners: Institut de Biologie de Lille (IBL) and CEA/iRTSB (Grenoble). The aim of the DOCK project, which stands for Docking at Grid, is to investigate parallel optimization metaheuristics for solving challenging conformational sampling and docking problems on the computational grid. This was the focus of a PhD thesis and a Post-doc position.

    • ANR CHOC (2006-2009). The CHOC project is a collaboration with the partners (OPALE, PRiSM/Versailles, MOAIS and GILCO / Grenoble). The new project dealt with the cooperation of metaheuristics and exact methods on grids for solving large problems. We were interested in the “academic" (or generic) problem of quadratic assignment (QAP, Q3AP). This was the focus of a PhD thesis and a Post-doc position which resulted in the coupling of two well-known optimization software platforms (ParadisEO and BOB++) and the resolution on Grid’5000 of difficult unsolved Q3AP instances (Best Paper Award in ACM GECCO’11).

    • ACI GRID Grid’5000 (2005-2008). Launched in 2003 by the government, the national Grid’5000 project aimed to create by 2007 a large computing testbed to promote scientific research on computational grids in France. It is an experimental testbed of 5000 processing cores belonging to several clusters distributed over 9 sites (Bordeaux, Grenoble, Nancy, Lille, Lyon, Orsay, Rennes, Sophia-Antipolis and Toulouse). As the Principal Investigator of the project, I received a budget of 200K€ start and lay the foundations of the site of Grid’5000 at Lille.

    • ACI GGM (2004-2007). GGM stands for Grille Géno-Médicale (Geno-Medical Grid). It is a collaboration project with two partners: PIRAMIDE from IRIT (Toulouse) and SIC from LIRIS (Lyon). The project aimed to propose a software architecture based on grids capable of managing heterogeneous and dynamic data in distributed biomedical data warehouses for intensive analysis and processing. The mission of our team consisted in studying the problems of combinatorial optimization on grids applied to knowledge discovery in distributed data warehouses.

    • ACI DOC-G (2001-2004). This project was a collaboration with the OPALE team (PRISM, UVSQ) and P2, O3 (ID-IMAG). The focus was put on the design and implementation of metaheuristics (our team) and exact methods (OPALE and O3 teams) on computaional grids. In this context, we coupled our ParadisEO platform with TakTuk (grid middleware developed by the O3 team). This was the focus of a PhD thesis.

  • National projects supported by Inria. I have been strongly involved in two major Inria national projects related to the Grid’5000 nation-wide grid infrastructure: Aladdin-G5K and Hemera, decribed in the following.

    • Aladdin-G5K (2007-2013). This is a nation-wide technological developement project involving all the sites of the Grid’5000 testbed. The overall objective of this project was to coordinate between the different sites of the testbed the activities in terms of system and network administration, the developement of the software tools allowing the exploitation of the testbed (resource reservation, system image deployment, resource supervision and monitoring, etc.). In terms of funding, Aladdin-G5K allowed me to hire two engineers for the system and network administration at Lille.

    • Hemera (2010-2014). This nation-wide project involving all the sites of the Grid’5000 testbed was the scientific counterpart of Aladdin-G5K. Within the framework of this project, I co-proposed a scientific challenge on large-scale combinatorial optimization3. Hemera enabled us to obtain from Inria a funding for a post-doc position in 2010 and a PhD thesis (2011-2014).

Hemera challenges: https://www.grid5000.fr/mediawiki/index.php/Hemera:SC:COPs

  • CPER Data (2015-2021, 530K€ for Grid’5000 equipment, 300K€ for 2 x 2-year engineer funding): in this regional project, that promotes research and software development related to advanced data science, I was the scientific leader of one of the three research lines of the project “Optimization and HPC”. In addition, the team is co-leader of the WP/lever “Research infrastructures” related to the Grid’5000. The team has managed the acquisition of a GPU-powered cluster integrated in Grid’5000 and the related scientific animation and training. The GPU-powered cluster was highly important for my research group and other AI and HPC teams. Moreover, I have contributed to hire an engineer (D. Delabroye) for its administration and related user support and development.

  • Industrial projects/collaborations. I have been involved mainly in two industrial contracts in collaboration with France Telecom R&D and Tasker Cloud Services:

    • France Telecom R&D (2003-2005). This contract was focused on the development of a software framework (DEMARNO) for the reusable design of hybrid parallel optimization methods for solving the multi-objective problem of mobile network design. This framework was part of a PhD thesis and has been extented later to optimization problems in other areas (molecular docking, vehicle routing, etc.).

    • Tasker Cloud Services (2011-2014). With the support of OSEO, this contract targeted the design and implementation of multi-objective combinatorial optimization methods for static and dynamic arbitration on multi-cloud infrastructures. This was the focus of a PhD thesis and several publications.

  • Other

I have been the Principal Investigator of the labeling project de of Université de Lille 1 as “Nvidia GPU Research Center" (2016).