everything.study
MATH 4806

Scientific Computing II (Honours)

0.5 credit · Undergraduate

Computational methods in numerical linear algebra including solvers for large linear systems, numerical optimization and eigenvalue problems. Application to topics such as scientific machine learning or parallel computing. Focus on implementation in an efficient compiled language and standard numerical libraries.

2021

Topics

  • Short Review: Linear Algebra, matlab
  • Compressed Row Storage for Sparse Matrices
  • Conjugate Gradient Method
  • Spectral Approximation (Power Method, Givens-Householder, Lanczos, ...)
  • How To: Use Numerical Linear Algebra Libraries (PETSc, LAPACK, ARPACK
  • Sparselib, ...)
  • Parallel Matrix-Vector Multiplication with MPI
  • Application: deep learning
  • Preconditioners

Prerequisites & course information

Also listed as COMP 4806 . Prerequisite(s): i) MATH 3806 ; and ii) MATH 2152 or MATH 2107 with a grade of C- or higher, or permission of the school. Lectures three hours a week.

Sources