PHY832 - International Course on Computational Physics (ICCP)

The dutch students are visiting during the week of Feb. 6th.
During this week we work on the first ICCP project, and you should spend
as much time as possible on the course and also some time socializing
with the dutch students while they are in East Lansing.

Instructor Phil Duxbury (BPS 4260, duxbury at pa.msu.edu)
TA Connor Glosser
Regular class time Tuesday 3-5pm BPS1240
Other times lab is open Tuesday and Thursday 10:20-11:40am

  • The course is in collaboration with Technical University Delft, in Delft the Netherlands. In the ICCP coding may be done in either C++ or Fortran 90, the choice being made by each group of students. Here are more details about the course structure.

    Course Outline
    This course involves developing Fortran (or C++ for the international option) code and using it to simulate several different systems of broad interest in physics. The main projects are based on the following methods / problems which form the basis of computational physics:
  • Monte Carlo Methods.
  • Molecular Dynamics Methods
  • Eigenvalue problems
  • Differential equations

    The projects

    Getting to know Fortran:
    To get started do the first two worksheets in the introduction to fortran course. There you will find a summary of Fortran syntax, some examples and a more comprehensive introduction to Fortran 90. A summary of linux commands is also there.

    Warm up project
    MC methods I and Outline of the first project
    MC methods II
    In the ICCP you should write your report in latex with the figures embedded in the text. Here is a template to use. mclectures1.tex This contains a figure and is called in such a way that the figure needs to be in the same directory as the .tex file. Here is the figure. harding2.jpeg. To create a .pdf file from the .tex file you only need to enter "pdflatex mclecture1.tex" to create a file called mclecture1.pdf. There are many good online introductions to latex.

    Course evaluation

    For each project you will write a report, which contains the objective of the calculations, your code, and the results you generated using your code. These reports constitute 75% of the grade. There will be a 1 hour final lab exam based on your projects. These meetings will be held in BPS1240. During your "final" meeting we will discuss your project reports and codes that you hand in, and you will be asked to run your algorithms to illustrate some issues. In addition you will be asked to make relatively minor modifications of your codes to calculate new things.

    Reference materials

  • PHY201 - introduction to fortran. This is the introductory Fortran course. The worksheets have some sample programs.
  • Fortran 90 reference card
  • Here is a more complete summary of Fortran 90 which we shall refer to during the course. It comes from the www site
  • Here is a list of useful unix commands and an introduction to Linux Computing in BPS1240

    Recommended text

  • Computational phyiscs, J.M. Thijssen (Cambridge University Press,1999)

    Other useful books

  • Introduction to computer simulations methods, Second Edition. H. Gould and J. Tobochnik (Addison-Wesley,1996)
  • Molecular modeling for beginners, A Hinchcliffe (Wiley,2003)
  • "An introduction to Fortran 90 for scientific computing", by James M. Ortega.
  • A more advanced book is
    "Fortran 90/95 explained" second edition. by Michael Metcalf and John Reid. Oxford University Press, 1999.
  • A nice (free) online book containing Fortran 77, Fortran 90 and C++ codes for a wide range of useful procedures is Numerical recipes online . They are charging for later editions.

    Some other useful links

    Fortran tutorial
  • Python reference card
  • C++ reference card
    Michael Feig's Lecture notes on biomolecular simulations using CHARMM. Read this to learn how to define the energy functions.
    Review of Monte Carlo Methods for proteins (.pdf file). Read this to learn how to choose the Monte Carlo moves.
    The rational behind force fields
    The CHARMM22 force field for proteins
    Review of simulation methods for macromolecules by Kurt Kremer (.pdf file)

    c++ or f90 - you make the call
    C++ is better for scientific computing
    f90 is better, including efficiency comparisons