Parallel scientific computing.
Instructor: Bastiaan J. Braams (braams@courant.nyu.edu).
Cross-listed as G22.2945-001 and G63.2012-001.
Class meets Thursdays 13:25-15:15, room 101, Warren Weaver Hall.
Prerequisites: Numerical Methods I (G22.2420/G63.2010) or Scientific Computing (G22.2112/G63.2043) or equivalent training in numerical analysis.
This is an advanced numerical analysis class, with emphasis on methods that are suitable for parallel scientific computing. Students will obtain experience in parallel computing using message passing and using threads. The class should be of interest to students in mathematics, computer science, or one of the applied sciences. Students must have good training in computer programming in a high level language. The class will require programming in C/C++ or Fortran. Prerequisites are Numerical Methods I (G22.2420/G63.2010) or Scientific Computing (G22.2112/G63.2043).
Text: Gregory R. Andrews, Foundations of Multithreaded, Parallel and Distributed Programming, Addison-Wesley, 2000.
Class notes 1: reading advice; C code for Laplace equation by point relaxation. Feb 8.