TOPICS IN NUMERICAL LINEAR ALGEBRA FOR HIGH-PERFORMANCE COMPUTING
As computer architectures evolve, numerical algorithms have to cope with high resolution simulations and data integration that are now key to many research fields. Solution methods for real world problems require a constantly increasingly computational effort, demanding for more and more resources and tailored algorithms.
In this talk we will introduce some recent high-performance algorithmic developments about solving dense linear systems, possibly numerical rank-deficient, and we will present some parallel techniques for the solution of large-scale sparse systems of linear equations.