Team : PEQUAN - Performance and Quality of Numerical Algorithms
Axes : SSR (👥👥), TMC (👥👥).Team leader :
Fabienne Jézéquel Campus Pierre et Marie Curie 26-00/309
No event planned at present.
Short presentation
The domain of the PEQUAN team deals with computer arithmetic and scientific computing. The team has developed the CADNA library based on a probabilistic approach of round-off error propagation. The skills of the team cover:
- the development of numerical algorithms using floating point or fixed point arithmetic and their numerical validation,
- the performance optimization of numerical applications on various architectures (multicore processors, GPUs, supercomputers,...),
- the problems related to image assimilation.
Computer arithmetic, Rounding errors, Numerical validation, Numerical quality, Heterogeneous and multimodal data, Distributed numerical computing, Linear algebra algorithms in mixed precision, Reliable results, Symbolic/numerical algorithms
Selected publications
- Th. Archambault, A. Filoche, A. Charantonis, D. Béréziat : “Multimodal Unsupervised Spatio-Temporal Interpolation of satellite ocean altimetry maps” Proceedings of the 18th International Conference on Computer Vision Theory and Applications, Lisboa, Portugal[Archambault 2023]
- S. Thiria, Ch. Sorror, Th. Archambault, A. Charantonis, D. Béréziat, C. Mejia, J.‑M. Molines, M. Crépon : “Downscaling of ocean fields by fusion of heterogeneous observations using Deep Learning algorithms” Ocean Modelling, vol. 182, pp. 102174, (Elsevier)[Thiria 2023]
- P. Amestoy, O. Boiteau, A. Buttari, M. Gerest, F. Jézéquel, J.‑Y. L'Excellent, Th. Mary : “Mixed Precision Low Rank Approximations and their Application to Block Low Rank LU Factorization” IMA Journal of Numerical Analysis, (Oxford University Press (OUP))[Amestoy 2022b]
- Th. Cameron, S. Graillat : “On a compensated Ehrlich-Aberth method for the accurate computation of all polynomial roots” Electronic Transactions on Numerical Analysis, vol. 55, pp. 401–423, (Kent State University Library)[Cameron 2022]
- M. Croci, M. Fasi, N. Higham, Th. Mary, M. Mikaitis : “Stochastic Rounding: Implementation, Error Analysis, and Applications” Royal Society Open Science, (The Royal Society)[Croci 2022]
- Q. Ferro, S. Graillat, Th. Hilaire, F. Jézéquel, B. Lewandowski : “Neural Network Precision Tuning Using Stochastic Arithmetic” NSV'22, 15th International Workshop on Numerical Software Verification,, Haifa, Israel[Ferro 2022b]
- A. Filoche, Th. Archambault, A. Charantonis, D. Béréziat : “Statistics-free interpolation of ocean observations with deep spatio-temporal prior” ECML/PKDD Workshop on Machine Learning for Earth Observation and Prediction (MACLEAN), Grenoble, France[Filoche 2022c]
- N. Higham, Th. Mary : “Mixed precision algorithms in numerical linear algebra” Acta Numerica, (Cambridge University Press (CUP))[Higham 2022]
- F. Lopez, Th. Mary : “Mixed Precision LU Factorization on GPU Tensor Cores: Reducing Data Movement and Memory Footprint” International Journal of High Performance Computing Applications, (SAGE Publications)[Lopez 2022]
- S. Graillat, F. Jézéquel, E. Queiros Martins, M. Spyropoulos : “Computing multiple roots of polynomials in stochastic arithmetic with Newton method and approximate GCD” International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2021), Timisoara, Romania[Graillat 2021]
Contact
Fabienne.Jezequel (at) nulllip6.fr