Séminaire CPU _ Yousef SAAD _« Parallel Multilevel Low-Rank approximation preconditioners »le 02 juin 2015 de 10h à 12h
Dans le cadre de l'axe 1 - "Calcul scientifique" - du cluster CPU, le Professeur Yousef SAAD (Université du Minnesota) donnera un séminaire intitulé «Parallel Multilevel Low-Rank approximation preconditioners ». Le séminaire aura lieu le mardi 2 Juin 2015, à 10h à l'INRIA (Salle Ada Lovelace).
This presentation will discuss a class of preconditioning methods for solving linear systems of equations that are based on exploiting low-rank approximations to certain matrices. These methods have a number of appealing features. Because they are essentially approximate inverse techniques, they handle indefiniteness quite well. Furthermore, they are amenable to SIMD compuations such those inherent to GPUs.
The talk will first describe a recursive divide and conquer approach geared toward Symmetric Positive Definite model problems issued from Finite Difference discretizations of PDEs. Then two extensions of this general approach will be described. The first exploits Schur complements in a parallel computing Domain Decomposition (DD) framework. The second extends this DD approach further by considering so-called `hierarchical interface decomposition further by considering so-called `hierarchical interface decomposition orderings' which are essentially algebraic generalizations of `wirebaskets' techniques used in Domain Decomposition methods.
Pour plus d'informations: