Logiciel
ParSys
Parallel Systems- ParSys
Research Themes
The team’s activity also involves the use of state-of-the-art parallel architectures to achieve optimal performance of codes developed for future exascale platforms. In addition, drawing inspiration from natural, efficient and robust phenomena, the team studies natural distributed algorithms to design efficient solutions for emerging networks, and even to develop robust distributed circuits in bacterial cell consortia, for computational or medical purposes (biological computers, smart drugs, etc.).
The main challenges concerning the algorithms studied in the ParSys team are: performance optimization, scaling and load balancing, fault tolerance and energy saving.
The applications (ranging from scientific computing, quantum computing and data analysis to network protocols and microbiological circuits) meet essential industrial or scientific needs and are the subject of industrial cooperation with ATOS-Bull, Total, and EdF, among others.
Coordination
Last publications
-
-
Pré-publication, Document de travail
Alexis Aune, Oguz Kaya. Latency hiding strategies for matrix multiplication on GPUs using pipelined multiword arithmetic. 2024. ⟨hal-04619183⟩
-
Thèse
Hugo Gabrielidis. High-performance machine learning and data analysis for next-generation railway design. Artificial Intelligence [cs.AIArtificial Intelligence]. Université Paris-Saclay, 2026. English. ⟨NNT : 2026UPAST019⟩. ⟨tel-05555019⟩
-
Thèse
Brice Chichereau. Optimization and evaluation of integrated HPC-Quantum software stacks. Computer Science [cs]. Université Paris-Saclay, 2026. English. ⟨NNT : 2026UPASG008⟩. ⟨tel-05551876⟩
-
Communication dans un congrès, Communication dans un congrès
Julien Rauch, Damien Rontani, Stéphane Vialle. Towards a Quantum Generative Graph-Based Clustering for Molecule Discovery. Quest-IS, Dec 2025, Palaiseau, France. pp.243-251, ⟨10.1007/978-3-032-13855-2_22⟩. ⟨hal-05549507⟩
-
Article dans une revue
Amanda Candemil Kanashiro, Hugo Gabrielidis, Filippo Gatti, Manoel Damião Sousa-Neto. Impact of artifact reduction using generative adversarial networks on diagnostic accuracy in cone-beam computed tomography. Journal of Dentistry, 2026, 167, pp.106400. ⟨10.1016/j.jdent.2026.106400⟩. ⟨hal-05549512⟩
-
Thèse
Atte Torri. Towards a fast task-based parallel tensor solver for high-dimensional problems. Numerical Analysis [cs.NA]. Université Paris-Saclay, 2025. English. ⟨NNT : 2025UPASG106⟩. ⟨tel-05534633⟩
-
Article dans une revue
Hugo Gabrielidis, Filippo Gatti, Stéphane Vialle. Physics-based super-resolved simulation of 3D elastic wave propagation adopting scalable diffusion transformer. Computer Physics Communications, 2026, 320, pp.109930. ⟨10.1016/j.cpc.2025.109930⟩. ⟨hal-05485963⟩
-
Communication dans un congrès
Joseph Touzet, Oguz Kaya, Pablo Arrighi, Amélia Durbec. QUIDS: A Large-Scale Distributed Framework for Quantum Irregular Dynamics Simulations. Q-CASA 2025 – IPDPS Workshop on Quantum Computing Algorithms, Systems, and Applications, Jun 2025, Milan, Italy. pp.491-500, ⟨10.1109/IPDPSW66978.2025.00080⟩. ⟨hal-05472605⟩
-
Communication dans un congrès
Brice Chichereau, Stéphane Vialle, Miwako Tsuji, Patrick Carribault, Mitsuhisa Sato. HPCQCMark: a new modular HPC-QC benchmarking framework. QCE – 2025 IEEE International Conference on Quantum Computing and Engineering, Aug 2025, Albuquerque, United States. pp.8-14, ⟨10.1109/QCE65121.2025.10285⟩. ⟨hal-05426530⟩