Numerical computation and data analysis are essential aspects of research in astrophysics and astronomy, a field which has historically always been a producer of big data. Thus, whether to analyze photometric or spectroscopic data from large observatories on the ground or in space, or to produce and analyze virtual data from numerical simulations, astrophysics tools require large computing resources and storage, as well as new numerical methods adapted to new or emerging architectures.
Theoretical predictions in astrophysics are largely based on complex numerical simulation codes which are high demands of massively parallel computing resources. These codes, which model various physical processes, are often highly multi-scale because of the structuring role of gravity, turbulence or scale couplings, and routinely need to model billions of resolution elements. Although the techniques differ from field to field, many of the physical processes modeled are common. Also, in a hardware landscape where hardware acceleration of the GPU type has taken off considerably and certainly promises to supplant “all-CPU” supercomputers, the methods developed a decade or more ago require a complete overhaul of their algorithms.
The deluge and complexity of data obtained by large surveys, multi-spectral strategies, or simulations, as well as the continuum of data infrastructures, have led to the emergence of new reduction and analysis tools. An increasingly significant part of these tools are based on artificial intelligence algorithms, possibly combined with more traditional Bayesian approaches. These most advanced developments in statistical analysis are still in their infancy, and a tremendous amount of work to educate the community about these powerful analytical tools remains to be done. These massive data being often dispersed at different levels of complexity on different computing and storage infrastructures, the problem then arises of the deployment of codes and the interfacing of the different devices.
This conference at the initiative of the Numerical Specific Action responds to the desire of the community to exchange their thoughts on these different numerical aspects in astrophysics. We encourage everyone to come and share their numerical analysis and modeling work, with the desire that this week of exchange will make it possible to decompartmentalize the themes, and to identify the methodological points of convergence between observations and simulations.
The conference will be held in Lyon (ENS Lyon) from December 12th to 16th, 2022.
Invited speakers
Mark Allen (CDS), Guillaume Aulanier (LPP), Clément Baruteau (IRAP), Emmanuel Bertin (IAP), Yann Capdeville (LPG) Benoît Cerutti (IPAG), Andrea Ciardi (LERMA), Arnaud Durocher (CEA), Damien Gratadour (LESIA), Mickael Grech (LULI), Patrick Hennebelle (CEA), Patrick Hudelot (IAP), Marc Huertas-Company (LERMA), Sylvie Joussaume (IPSL), Laurène Jouve (IRAP), François Lanusse (CEA), Geoffroy Lesur (IPAG), Aurélie Marchaudon (IRAP), Patrick Masse (AMD), Shan Mignot (OCA), Jérôme Pety (IRAM), Alejandra Recio-Blanco (OCA), Stéphane Requena (GENCI), Jenny Sorce (CRIStAL), Jean-Pierre Vilotte (INSU)
Registration
For your convenience, we will suggest that you organize lunches at the local CROUS restaurant.We will also organize a conference dinner on Wednesday December 14, in the restaurant "Bouchon Bât d'Argent".
The lunches and conference dinner will be free of charge for all the registered particpants.
A welcom cocktail will be offered at the end of the talks on Monday 12th.
Loading... 
