Weekly update October 23-27

Hi all, project 4 is now availbale and we will discuss it tomorrow and Friday during the lectures. The topic is the simulation of phase transitions for magnetic systems using a binary model, that is the interacting object can only take two values. The model is the so-called Ising model, widely used in many scientific applications, from statistical studies to modeling of phase transitions in chemistry and physics. 

The model is described in chapter 13 of the lecture notes and we will start with discussing some simple properties and how to implement the model numerically.  We will simulate this system using the famous Metropolis algorithm, which is a special implementation of a so-called Markov chain. Markov chains and the Metropolis algorithm will also be discussed during Thursday's and Friday's lectures. Chapter 12 deals with random walks, Markov chains and the Metropolis algorithm. But we start on Thursday with a discussion of project 4.

This project is an example of a project where we will need quite som CPU time and parallelization comes in handy. We will discuss parallelization with examples next week at the lab (first hour each session) and we will focus on two ways to parallelize, either using MPI or OpenMP. These are described in the slides on parallelization, see http://compphysics.github.io/ComputationalPhysics/doc/pub/codeoptimization/html/codeoptimization.html

Else, we (all five) have been procrastinating the feedback on  project 2. We are very sorry for this and will try to fix asap!  The deadline for project 3 is Friday.   I take also the liberty of reminding you on how we grade projects 3-5, see https://github.com/CompPhysics/ComputationalPhysics/blob/master/doc/Projects/ProjectExample/projecteval.txt

A git pull gives you the latest update on project 4. 

Best wishes to you all,

Morten