Recent selected talks

• An overview of mixed precision strategies for scientific computing (NGT - Openlab "Optimising Floating Point Precision" Workshop , CERN, Geneva, July 2025). Video of the talk.
• Mixed precision algorithms in numerical linear algebra (MUMPS workshop , CINES, Montpellier, June 2025).
• Mixed precision preconditioned GMRES (Sparse Days , Toulouse, June 2025).
• Mixed precision accumulation for neural network inference guided by componentwise forward error analysis (PEQUAN team meeting, Paris, April 2025).
• Condense & Distill: fast distillation of large floating-point sums via condensation (French--Japanese Workshop on Numerical Computing, Paris, March 2025).
• Nineteen dubious ways to compute low-rank approximations in mixed precision (Seminar at LIP, Lyon, November 2024).
• Optimal quantization of rank-one matrices for butterfly factorizations (SIAM LA24, Paris, May 2024).

Older talks

• Adaptive Precision Solvers and Preconditioners (Advances in Numerical Linear Algebra: Celebrating the 60th Birthday of Nick Higham, Manchester, July 2022). Video of the talk.
• Adaptive Precision Solvers for Sparse and Data Sparse Sytems (Sparse Days, Saint-Girons, June 2022).
• Reducing Data Movement in Mixed Precision LU Factorization with GPU Tensor Cores (SIAM PP22, online, February 2022).
• Data-Aware Mixed Precision Algorithms (SIAM AN21, online, July 2021).
• Are numerical algorithms accurate at large scale and at low precisions ? (20th IMA Leslie Fox Prize Event, online, June 2021). Video of the talk.
• Opportunities for Mixed Precision Arithmetic in Numerical Linear Algebra (ARITH21, online, June 2021). Video of the talk.
• Exploiting Mixed Precision Arithmetic in the Solution of Linear Systems (SIAG/LA Early Career Prize lecture at SIAM LA21, online, May 2021).
• Multiple Word Arithmetic with GPU Tensor Cores: Theory and Practice (SIAM LA21, online, May 2021).
• Mixed Precision Low Rank Compression and its Application to BLR Matrix Factorization (SIAM CSE21, online, March 2021).
• Mixed Precision Low Rank Compression of Data Sparse Matrices (Communications in NLA, online, September 2020). Video of the talk.
• Performance and Accuracy of Mixed-Precision Matrix Factorizations with GPU Tensor Cores (SIAM PP20, Seattle, February 2020).
• Asymptotic Complexity of Low-rank Sparse Direct Solvers with Sparse Right-hand Sides (SIAM CSC20, Seattle, February 2020).
• A New Approach to Probabilistic Rounding Error Analysis (SIAM CSE19, Spokane, February 2019).
• Sharper and Smaller Error Bounds for Low Precision Scientific Computing (Seminar in AriC team, Lyon, Oct 2019).
• Exploiting Fast Matrix Arithmetic in Block Low-Rank Factorizations (Structured Matrix Days, Limoges, May 2019).
• Accuracy and Stability of Block Low-Rank Linear Solvers (Seminar in PEQUAN team, Paris, December 2018).
• Complexity, Scalability, and Stability of Block Low-Rank Solvers (Seminar at Laboratoire d'Informatique du Parallélisme, Lyon, December 2018).
• Accuracy and Stability of Low-Rank Linear Solvers (Seminar at Inria Bordeaux, November 2018).
• A New Preconditioner for Low-Accuracy Block Low-Rank Multifrontal Solvers (Sparse Days, Toulouse, September 2018).
• Accelerating Linear Systems Solution by Exploiting Low-Rank Approximations to Factorization Error (IMA Conference on NLA and Optimization, Birmingham, June 2018).
• Complexity and Parallelism of the Solution Phase in Sparse Direct Solvers (PMAA18, Zürich, June 2018, by Gilles Moreau).
• Bridging the Gap between Flat and Hierarchical Low-Rank Matrix Formats (Structured Matrix Days, Lyon, May 2018).
• Distributed-Memory BLR Factorization for Large-Scale Systems and Applications (SIAM PP18, Tokyo, March 2018).
• Block Low-Rank Matrices: Main Results and Recent Advances (Seminar at Rutherford Appleton Laboratory, November 2018).
• Block Low-Rank Matrices: Main Results and Recent Advances (Seminar at Laboratoire Jean Kuntzmann, Grenoble, July 2018).
• Block Low-Rank Multifrontal Solvers: Complexity, Performance, and Scalability (My PhD defense, Toulouse, 24 November 2017).
• Block Low-Rank Multifrontal Sparse Direct Solvers (MATHIAS, October 2017, Paris).
• Block Low-Rank Multifrontal Solvers: Complexity, Performance, and Scalability (Sparse Days, Toulouse, September 2017).
• Large-Scale 3D CSEM Modeling with a BLR MUMPS Solver (SIAM GS17, Erlangen, September 2017, by Daniil Shantsev).
• On the Comparison of Sparse Multifrontal Hierarchical and Block Low-Rank Solvers (MUMPS User Days, Montbonnot St-Martin, June 2017).
• Comparison of BLR and HSS Low-Rank Formats in Multifrontal Solvers: Theory and Practice (SIAM CSE17, Atlanta, February 2017).
• Multicore Performance of the Block Low-Rank Multifrontal Factorization (Journée Lyon Calcul, Lyon, December 2016).
• Performance and Scalability of the Block Low-Rank Multifrontal Factorization (PhD Days at ENSEEIHT, Toulouse, November 2016).
• Performance and Scalability of the Block Low-Rank Multifrontal Factorization (PMAA16, Bordeaux, July 2016).
• On the Complexity of the Block Low-Rank Multifrontal Factorization (Sparse Days, Toulouse, June 2016).
• Complexity and Performance of the Block Low-Rank Multifrontal Factorization and its Variants (SIAM PP16, Paris, April 2016).
• A Comparison of Parallel Rank-Structured Solvers (SIAM PP16, Paris, April 2016, by François-Henry Rouet).
• Sparse Direct Solvers towards Seismic Imaging of Large 3D Domains (EAGE16, Vienna, 2016, by Patrick Amestoy).
• Complexity and Performance of Block Low-Rank Multifrontal Solvers (Journée des Doctorants, Université Paul Sabatier, Toulouse, 2015).
• 3D Frequency-Domain Seismic Modeling with a Parallel BLR Multifrontal Direct Solver (PhD Days at ENSEEIHT, Toulouse, November 2015).
• Performance of Random Sampling for Computing Low-rank Approximations of a Dense Matrix on GPUs (SC15, Austin, November 2015, by Piotr Luszczek).
• 3D Frequency-Domain Seismic Modeling with a Parallel BLR Multifrontal Direct Solver (SEG15, New Orleans, October 2015).
• A Comparison of Different Low-Rank Approximation Techniques (SIAM LA15, Atlanta, October 2015, by François-Henry Rouet).
• Improving Multifrontal Solvers by means of Block Low-Rank Approximations (Workshop on fast solvers, Toulouse, June 2015).
• 3D Frequency-Domain Seismic Modeling with a Parallel BLR Multifrontal Direct Solver (SEISCOPE Annual Meeting, Grenoble, May 2015).
• Improving Multifrontal Solvers by means of Block Low-Rank Approximations (Workshop at Livermore Software Technology Corporation, March 2015).
• Performance of Random Sampling for Computing Low-rank Approximations of a Dense Matrix on GPUs (SIAM CSE15, Salt Lake City, March 2015).
• Performance of Computing Low-Rank Matrix Approximation on a Hybrid CPU/GPU Architecture (SIAM CSE15, Salt Lake City, March 2015, by Ichitaro Yamazaki).