This poster describes some activities developed by CNR and its third party within EoCoE

Computing eigenvectors of graph Laplacian is a main computational kernel in data clustering, i.e., in identifying different groups such that data in the same group are similar and points in different groups are dissimilar with respect to a given notion of similarity. Data clustering can be reformulated in terms of a graph partitioning problem when the given set of data is represented as a graph, also known as similarity graph.

In this paper we revisit the problem of performing a QR-step on an unreduced Hessenberg matrix H when we know an "exact" eigenvalue ?0 of H. Under exact arithmetic, this eigenvalue will appear on diagonal of the transformed Hessenberg matrix H~ and will be decoupled from the remaining part of the Hessenberg matrix, thus resulting in a deflation. But it is well known that in finite precision arithmetic the so-called perfect shift can get blurred and that the eigenvalue ?0 can then not be deflated and/or is perturbed significantly.

Modal decomposition techniques are used to analyse the wake field past a marine propeller achieved by previous numerical simulations (Muscari et al. Comput. Fluids, vol. 73, 2013, pp. 65-79). In particular, proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD) are used to identify the most energetic modes and those that play a dominant role in the inception of the destabilization mechanisms. Two different operating conditions, representative of light and high loading conditions, are considered.

A genuinely two-dimensional discretization of general drift-diffusion (including incompressible Navier--Stokes) equations is proposed. Its numerical fluxes are derived by computing the radial derivatives of "bubbles" which are deduced from available discrete data by exploiting the stationary Dirichlet--Green function of the convection-diffusion operator. These fluxes are reminiscent of Scharfetter and Gummel's in the sense that they contain modified Bessel functions which allow one to pass smoothly from diffusive to drift-dominating regimes.

The generalized Schur algorithm is a powerful tool allowing to compute classical decompositions of matrices, such as the QR and LU factorizations. When applied to matrices with particular structures, the generalized Schur algorithm computes these factorizations with a complexity of one order of magnitude less than that of classical algorithms based on Householder or elementary transformations.