Code and final report for the 2 projects part of the class HPC for numerical methods and data analysis taught by Laura Grigori at EPFL.
- The first project was individual. It compares 3 different algorithms (Classical Gram-Schmidt, Cholesky-QR and TSQR) which compute the thin QR factorization of a matrix.
- The second project was implemented in a group of 2 students. It investigates the numerical stability and parallel performance of the randomized Nyström algorithm used to compute a rank-k approximation of a square positive semi-definite matrix.