Searcharxiv⌕ Search

arXiv subjects

Théo Mary

Publications and source records attributed to Théo Mary.

2 recordsLinked to original sources

3D Modeling of Solar Oscillations with Hybridizable Discontinuous Galerkin Method

With increasing quantity and quality of solar observations, it becomes essential to account for three-dimensional heterogeneities in wave modeling for seismic data interpretation. In this context, we present a 3D solver of the time-harmonic adiabatic stellar oscillation equations without background flows on a domain consisting of the Sun and its photosphere. The background medium consists of 3D heterogeneities on top of a radial strongly-stratified standard solar model. The oscillation equations are solved with the Hybridizable Discontinuous Galerkin (HDG) method, considering a first-order formulation in terms of the vector displacement and the pressure perturbation. This method combines the high-order accuracy and the parallelism of DG methods while yielding smaller linear systems. These are solved with a direct solver, with block low-rank compression and mixed-precision arithmetic to reduce memory footprint. The trade-off between compression and solution accuracy is investigated, and our 3D solver is validated by comparing with resolution under axial symmetry for solar backgrounds. The capacity of the solver is illustrated with wave speed heterogeneities characteristic of two physical phenomena: active regions and convection. We show the importance of global 3D gravito-acoustic wave simulations, in particular when the amplitudes of the perturbations are strong and their effect on the wavefield cannot be estimated by linear approximations.

math.AP↗

Matrix-free construction of HSS representation using adaptive randomized sampling

We present new algorithms for the randomized construction of hierarchically semi-separable matrices, addressing several practical issues. The HSS construction algorithms use a partially matrix-free, adaptive randomized projection scheme to determine the maximum off-diagonal block rank. We develop both relative and absolute stopping criteria to determine the minimum dimension of the random projection matrix that is sufficient for the desired accuracy. Two strategies are discussed to adaptively enlarge the random sample matrix: repeated doubling of the number of random vectors, and iteratively incrementing the number of random vectors by a fixed number. The relative and absolute stopping criteria are based on probabilistic bounds for the Frobenius norm of the random projection of the Hankel blocks of the input matrix. We discuss parallel implementation and computation and communication cost of both variants. Parallel numerical results for a range of applications, including boundary element method matrices and quantum chemistry Toeplitz matrices, show the effectiveness, scalability and numerical robustness of the proposed algorithms.

math.NA↗