Pierre Kestener

Atomic Energy and Alternative Energies Commission | CEA · IRFU - Institut de Recherche sur les lois Fondamentales de l'Univers

Aims. Clouds are expected to form in a broad range of conditions in the atmosphere of exoplanets given the variety of possible condensible species. This diversity, however, might lead to very different small-scale dynamics depending on radiative transfer in various thermal conditions. Here, we aim to provide some insight into these dynamical regime...

Aims. We investigate the impact of dimensionality, resolution, and long timescales on convective numerical simulations forced by thermo-compositional diabatic processes. We focus our study on simulations that are stable to the Schwarzschild criterion but unstable to the Ledoux one (i.e. simulations with a stabilising temperature gradient and a dest...

Clouds are expected to form in a wide range of conditions in the atmosphere of exoplanets given the large range of possible condensible species. However this diversity might lead to very different small-scale dynamics depending on radiative transfer in various thermal conditions: we aim at providing some insights into these dynamical regimes. We pe...

We investigate the impact on convective numerical simulations of thermo-compositional diabatic processes. We focus our study on simulations with a stabilizing temperature gradient and a destabilizing mean-molecular weight gradient. We aim to establish the possibility for a reduced temperature-gradient in such setups. A suite of 3D simulations were...

Parallel programming models overview (OpenMP, OpenACC, OpenMP target, Kokkos, SYCL)

Numerical codes using the lattice Boltzmann methods (LBM) for simulating one-or two-phase flows are widely compiled and run on graphical process units. However, those computational units necessitate to rewrite the program by using a low-level language which is suited to those architectures (e.g. CUDA for GPU NVIDIA or OpenCL). In this paper we focu...

Numerical codes using the Lattice Boltzmann Methods (LBM) for simulating one- or two-phase flows are widely compiled and run on graphical process units. However, those computational units necessitate to re-write the program by using a low-level language which is suited to those architectures (e.g. CUDA for GPU NVIDIA or OpenCL). In this paper we fo...

Presentation about a work in progress consisting in refactoring some core algorithms involved in Non-Destructive Testing (NDT) using C++/Kokkos library for performance portability. We first review what performance portability is and what are the others software library and programing models addressing this issue. We explain why we chose to use Kok...

It is still not known whether solutions to the Navier-Stokes equation can develop singularities from regular initial conditions. In particular, a classical and unsolved problem is to prove that the velocity field is Hölder continuous with some exponent h<1 (i.e., not necessarily differentiable) at small scales. Different methods have already been p...

By generalizing the theory of convection to any type of thermal and compositional source terms (diabatic processes), we show that thermohaline convection in Earth's oceans, fingering convection in stellar atmospheres, and moist convection in Earth's atmosphere are derived from the same general diabatic convective instability. We also show that "rad...

Report about using C++/kokkos library for high-performance CFD applications

Conference: 2019 PATC Training - Performance portability for GPU application using high-level programming approaches Introduction to using Kokkos/C++ library (OpenMP and CUDA backend) for designing performance portable applications.

Convection is an important physical process in astrophysics well-studied using numerical simulations under the Boussinesq and/or anelastic approximations. However these approaches reach their limits when compressible effects are important in the high Mach flow regime, e.g. in stellar atmospheres or in the presence of accretion shocks. In order to t...

By generalizing the theory of convection to any type of thermal and compositional source terms (diabatic processes), we show that thermohaline convection in Earth oceans, fingering convection in stellar atmospheres, and moist convection in Earth atmosphere are deriving from the same general diabatic convective instability. We show also that "radiat...

Revisiting LBM algorithms and numerical implementations and optimisations using the Kokkos library. Performance measurements on multiple architectures (Intel Skylake, Intel KNL, ARM ThunderX2, Nvidia P100). Applications to multiphase flow by coupling LBM Navier-Stkokes with phase field model (Allen-Cahn).

Using the kokkos library to obtained parallel implementation of two families of high-order numerical schemes (MOOD and Spectral Difference Methods). Applications to compressible Euler hydrodynamics. Performance portable Kokkos library allows us to write generic implementation and test on multiple hardware architectures (Intel Skylake, Intel KNL, Nv...

- a short overview of exascale hardware trends - overview of performance portability - LBM (Lattice Boltzmann) kernels implemented with C++/kokkos library - LBM kernels optmizations and performance measuments on multiple architectures (Skylake, KNL, Nvidia GPUs, ...)

- Present some hardware trends in HPC - stress the need to use high level programing approach to design parallel software application - illustrate performance portability solutions by the use of c++ kokkos library to design high order numerical schemes (spectral difference method) for CPU, GPU, ARM .... - short presentation of canop, adaptive mesh...

Evaluation of a performance portability library called Kokkos on mini-applications taken from the CIVA non-destructive computation (NDT) simulation tool developed by the CEA (French Atomic Energy Commission). A related paper will be submitted soon to ParCo 2019 conference.

This contribution deals with the fluid modeling of multicomponent magnetized plasmas in thermo-chemical non-equilibrium from the partially- to fully-ionized collisional regimes, aiming at the predictive simulation of magnetic reconnection in Sun chromosphere conditions. Such fluid models are required for large-scale simulations by relying on high p...

The present contribution starts from a two-temperature single-momentum multicom-ponent diffusion model coupled to Maxwell's equations obtained in Wargnier et al. (2018a). It is the asymptotic limit of small Debye length of the model rigorously derived from kinetic theory by Graille et al. (2009) using a multi-scale Chapman Enskog expansion. The mod...

Rapid overview of Spectral Difference Methods implemented with C++/Kokkos library

Slides presented during PATC training on high-level programing for performance portability, 31st May and June 1st 2018. https://events.prace-ri.eu/event/712/

Introduction to using Kokkos library (OpenMP and CUDA backend) for designing performance portable application.

Broad and general introduction to High Performance Computing

Broad and general introduction to GPU computing for students at Ecole Doctorale ED127 (Astrophysique - Ile de France)

Exascale implies a major pre-requisite in terms of energy efficiency, as an improvement of an order of magnitude must be reached in order to stay within an acceptable envelope of 20 MW. To address this objective and to continue to sustain performance, HPC architectures have to become denser, embedding many-core processors (to several hundreds of co...

We’ll report on the use of the kokkos C++ library for designing new performance portable implementations of the algorithms used in astrophysics computational fluid dynamics applications. Among others libraries with similar features, kokkos, which is developed at Sandia National Laboratory, provides a very promising way of designing high-performance...

Magnetohydrodynamic (MHD) turbulence driven by the magnetorotational instability can provide diffusive transport of angular momentum in astrophysical disks, and a widely studied computational model for this process is the ideal, stratified, isothermal shearing box. Here we report results of a convergence study of such boxes up to a resolution of $N...

Slides presenting the Kokkos library developped at SandiaNL, illustrating performance portability concepts in c++ in the context of high performance computing applications.

A very short overview of high-level parallel programing models for (performance) portability: directives (OpenMP/OpenACC), high-level librairies (Kokkos, RAJA, arrayFire, hpx, ...) or DSL.

Broad and general introduction to parallel programing with MPI (Message Passing Interface)

A Software package providing a wavelet-based multifractal analysis tool implementing the WTMM (Wavelet Trasform Modulus Maxima) method. http://www.scholarpedia.org/article/Wavelet-based_multifractal_analysis

Presents an overview of some recent work on developping an applicative software layer on top of p4est, to ease starting new application requiring adaptive mesh refinement for solving PDE problems (CFD, astrophysics, ...)

This work focuses on the numerical approximation of the Shallow Water Equations (SWE) using a Lagrange-Projection type approach. We propose to extend to this context recent implicit-explicit schemes developed in the framework of compressibleflows, with or without stiff source terms. These methods enable the use of time steps that are no longer cons...

Many physical problems involve spatial and temporal inhomogeneities that require a very fine discretization in order to be accurately simulated. Using an adaptive mesh, a high level of resolution is used in the appropriate areas while keeping a coarse mesh elsewhere. This idea allows to save time and computations, but represents a challenge for dis...

A short overview of existing adaptive mesh refinement codes. Introduction to code CanoP (designed on top of p4est library).

Présentation faite à la 3ème journée Mésochallenges de GENCI du 7 octobre 2015 ( (Institut Poincaré)

Présentation faite à la 3ème journée Mésochallenges de GENCI du 7 octobre 2015 ( (Institut Poincaré)

Some slides presented at the INFIERI 2014 summer school in Paris (July 2014) : http://infieri2014summerschool.in2p3.fr/

Theses are some slides presented at our PATC (PRACE Advanced Training Center) training event "Introduction to GPU programming" given at Maison de la Simulation, December 2014.

A mini cython + gpu tutorial given at Maison de la Simulation (pizzaminar - Dec 3, 2014)

Maison de la Simulation, internal seminar. 3 subjects discussed: 1. presents parallel IO performance mesurements using code RamsesGPU on a very simulation on TITAN. 2. Some basic considerations about discrete time and continuous time Fourier transform as well as small presentation of existing libraries for parallel distributed FFT. 3. a very short...

The Scientific Trigger Unit (UTS) is a satellite equipment designed to detect Gamma Ray Bursts (GRBs) observed by the onboard 6400 pixels camera ECLAIRs. It is foreseen to equip the low-Earth orbit French-Chinese satellite SVOM and acts as the GRB trigger unit for the mission. The UTS analyses in real-time and in great details the onboard camera da...

The on-board Scientific Trigger Unit (UTS) is designed to detect Gamma Ray Bursts (GRBs) in real-time, using the data produced by the ECLAIRs camera, foreseen to equip the future French-Chinese satellite mission SVOM (Space-based Variable Objects Monitor). The UTS produces GRB alerts, sent to the ground for GRB follow-up observations, and requests...

The 2D Wavelet-Transform Modulus Maxima (WTMM) method was used to detect microcalcifications (MC) in human breast tissue seen in mammograms and to characterize the fractal geometry of benign and malignant MC clusters. This was done in the context of a preliminary analysis of a small dataset, via a novel way to partition the wavelet-transform space-...

The main objective of the COMPASS project is to provide a full scale end-to-end AO development platform, able to address the E-ELT scale and designed as a free, open source numerical tool with a long term maintenance plan. The development of this platform is based on a full integration of software with hardware and relies on an optimized implementa...

A full-time beam-former for two independent antenna groups, with visibility computation capabilities at a slower rate, has been formerly designed on a single FPGA for the BAO-radio instrument, a radio telescope demonstrator for the study of dark energy by the observation of Baryonic Acoustic Oscillations (BAO) spectrum with HI probing technique. On...

The ANTARES Neutrino Telescope was completed in May 2008 and is the first operational Neutrino Telescope in the Mediterranean Sea. The main purpose of the detector is to perform neutrino astronomy and the apparatus also offers facilities for marine and Earth sciences. This paper describes the design, the construction and the installation of the tel...

The BAO radio instrument is a demonstrator of a radio telescope dedicated to cosmological studies and dark energy observation. It will analyze the signal emitted by the hydrogen in the universe with the aim of the detection of BAO's (Baryon Acoustic Oscillations). The instrument will use cylindrical reflectors or a large number of small dishes asso...

Finite volume numerical methods have been widely studied, implemented and parallelized on multiprocessor systems or on clusters. Modern graphics processing units (GPU) provide architectures and new programing models that enable to harness their large processing power and to design computational fluid dynamics simulations at both high performance an...

The multifractal nature of solar photospheric magnetic structures are studied using the 2D wavelet transform modulus maxima (WTMM) method. This relies on computing partition functions from the wavelet transform skeleton defined by the WTMM method. This skeleton provides an adaptive space-scale partition of the fractal distribution under study, from...

We propose and test a wavelet transform modulus maxima method for the au- tomated detection and extraction of coronal loops in extreme ultraviolet images of the solar corona. This method decomposes an image into a number of size scales and tracks enhanced power along each ridge corresponding to a coronal loop at each scale. We compare the results a...

Magnetohydrodynamic turbulence is thought to be responsible for producing complex, multiscale magnetic field distributions in solar active regions. Here we explore the multiscale properties of a number of evolving active regions using magnetograms from the Michelson Doppler Imager (MDI) on the Solar and Heliospheric Observatory (SOHO). The multifra...

A 2D wavelet transform modulus maxima (WTMM) method is used to characterise the complexity of the distribution of the photospheric magnetic field of active regions. The WTMM method offers increased accuracy and reliability over previous fractal and multifractal methods. The multifractal spectrum of both quiet Sun and active region magnetic features...

The 2D wavelet transform modulus maxima method is used to obtain quantitative information on the fractal/multifractal nature and anisotropic structure of Galactic atomic hydrogen (H I) from the Canadian Galactic Plane Survey. Five mosaics were analyzed in the second quadrant of the Milky Way, corresponding to the Local, Perseus, and Outer spiral ar...

Motivations Qu'est ce que l'Inpainting Exemples d'utilisation de l'Inpainting L'Inpainting utilisé en astrophysique Implantation de bases / dictionnaires pour l'Inpainting en traitement d'images La Transformée en Ondelettes Discrète par Lifting scheme Implantation de la TO 2D CDF dans CUDA La Transformée en Cosinus Discrète 2D (DCT) Implantation de...

http://www.scholarpedia.org/article/Wavelet-based_multifractal_analysis

Gamma-ray bursts (GRB) sign energetic explosions in the Universe, occurring at cosmological distances. Multi-wavelength observations of GRB allow to study their properties and to use them as cosmological tools. In 2012 the space borne gamma-ray telescope ECLAIRs is expected to provide accurate GRB localizations on the sky in near real-time, necessa...

Presentation made in June 2007 at HSHS workshop about some development made for BAO radio prototype equipment (designed to detect Baryonic Acoustics Oscillations). The contribution here was to study the possibility to design a highrate FFT-based operator on a distributed cluster of FPGA (at the heart of the correlator).

Simply explain a common problem in spectral analysis : spectral leakage and brievely introduce the concept of polyphase filter bank designed to avoid this problem.