Weekly plans and update for week 37

Dear All and welcome to a new week and thx a million for heroic efforts on project 1. Project 2 is now out and a git pull gives you the latest exciting project!  It deals with eigenvalue problems and brings us over to this week's new topic, linear algebra and eigenvalue solvers. The material is covered by chapter 7 of the lecture notes and slides at http://compphysics.github.io/ComputationalPhysics/doc/pub/eigvalues/html/eigvalues.html

 

We will however start Thursday's lecture with a brief summary from last week, with a short repetition on LU decomposition. We will finalize the discussion on iterative linear equation solvers and move on to interpolation methods before we start with eigenvalue methods. Iterative solvers and interpolation methods like spline methods are discussed in the linear algebra slides and chapter 6 of the lecture notes. Cubic spline, one of the most popular interpolation methods utilizes a tridiagonal solver like the one you developed for project 1. It means that with project 1 you could write your own interpolation code. Pretty cool. 

 

Else, at the lab this week we will discuss how to speed up our code with minimal efforts, gaining a speed-up of a factor of three almost. Black magic? No, vectorization, and it is described under the slides at http://compphysics.github.io/ComputationalPhysics/doc/pub/codeoptimization/html/codeoptimization.html

 

 

Best wishes to you all from the lab+lecture gang

ADHJM

 

Publisert 11. sep. 2019 21:07 - Sist endret 11. sep. 2019 21:07