Physics 780.20 "Computational Physics"
Winter, 2009

General Information about 780.20 Computational Physics

Course title:

Computational Physics

References:

There is no required text but there will be readings for each class session from handouts passed out in class and background notes posted online. Some useful references that we'll consult are:

• The 2008 lecture notes [pdf] by Morten Hjorth-Jensen from the University of Oslo. Prof. Hjorth-Jensen's philosophy of teaching computational physics is similar to mine and he covers similar topics. (His course web pages: FYS3150 and FYS4410.)
• Computational Physics: Problem Solving with Computers by Rubin Landau and Manuel Paez is the text we've used in the past. It is not required but is a useful guide, including as a good source of projects. Many of the codes we'll explore originated from this book. [It is available as an "E-book" from the OSU library---you can view any part in your browser but you cannot print more than a page at a time.] An updated version called A Survey of Computational Physics is available as well.
• I've recently discovered Writing Scientific Software: A Guide to Good Style by Oliveira and Stewart, which is available from Amazon.com for only $10 used.
• For more about C++ object-oriented programming, look at A First Course in Computational Physics and Object-Oriented Programming with C++ by David Yevick [contents and excerpts are available on Amazon.com] and C++ and Object-oriented Numeric Computing for Scientists and Engineers by Daoqi Yang.
• Numerical Recipes: The Art of Scientific Computing by Press et al. (2nd edtion) is available in online editions.
• Python Scripting for Computational Science by Langtangen is a good reference for the basics of Python and what you need to know to do computational physics with it. The link is to the OSU E-book version, from which you can get PDF's of individual chapters.
• There are many good C++ references to choose from if you want to supplement the class. Here are a few that are available online for OSU people from the Safari page.
  • Teach Yourself C++ in 21 Days by J. Liberty and B.L. Jones. The title sounds a bit goofy but the order of topics and the pacing is well suited for learning C++ quickly.
  • C++ Primer Plus by Stephen Prata. This has very good reviews, but works better as a reference than a short-term learning guide.
  • Practical C++ Programming by Steve Oualline (published by O'Reilly). This has a lot of good advice for writing C++ programs.
• Another excellent resource for C++ programming is http://www.cplusplus.com (if you Google a C++ command, this is likely to be the first hit).

Prerequisites:

The prerequisites are simply physics at least through the undergraduate 26x series (although some have successfully taken the course concurrently with 262!). It will be useful but not necessary to have some experience with Mathematica, MATLAB, Python, C, fortran, or C++. The teaching strategy is to give you computer programs and have you run and then modify (or debug) them as you follow along through worksheets. Email or visit Prof. Furnstahl if you're concerned about your preparation (e.g., if you have no experience at all).

Material:

We'll start with an overview based on the first part of the Hjorth-Jensen lecture notes and then cover selections from the rest of the notes plus topics based on the Landau/Paez text and on the instructors' latest prejudices and class interest (to be determined!). In most cases the discussion will be framed by a physics topic such as nonlinear oscillations (e.g., chaos). We'll be using programs written in C++ and Python and occasionally Matlab (or Mathematica) as we go along. Some topics we will cover along the way:

• Errors and uncertainties in computations. E.g., one should understand how to analyze whether a calculation is limited by the algorithm or round-off error. We will come back to this topic repeatedly.
• Basic computational algorithms for: integration, differentiation, differential equations, root finding. Less emphasis on theory than on understanding how well an algorithm should work (e.g., should the accuracy improve as 1/N², where N is the number of points used and does it?) and what algorithm is appropriate for what situation (e.g., oscillatory integrals or integrands with singularities). In many (or most) cases you should be using a packaged library routine and not writing your own, so we'll learn how to use such a library and check the results.
• What you should know about: random numbers, Monte Carlo integration and simulation, matrix computing, calling Fortran libraries from C++, plus additional topics as time permits.
• Aspects of writing code: good programming practices; how to test and debug a code (C++, fortran, MATLAB, or whatever); how to tune a code to run faster.
• Aspects of a computational physics project: breaking down a project into sub-problems; implementation issues (e.g., program design, code conventions, makefiles); use of graphics for visualization; verification.
• Object-oriented programming: What is it and when is it relevant for computational problems?
• Using Mathematica or MATLAB for computational physics. This is a broad topic, of course, and we will just touch upon aspects here.

Computing Environment:

The general idea is to use basic and portable tools. The homepage will have details about setting these up on your personal computer.

• The computers in Smith 1094 can be run with Linux or Windows XP. You can choose which to use.
• For Linux users, the computer environment include the GNU tools (also available in Smith 1094). These include g++, make, indent, gdb, gprof, and editors (e.g., emacs, nedit).
• For Windows users, the computing environment will be mainly Cygwin, which simulates the GNU/Linux environment. (You can also log into a public Linux machine via an X-windows program, Xwin32.) Sometimes we will use the Dev-C++ IDE.
• We'll have the INTEL compilers (for C++ and Fortran 90/95) available on both platforms.
• We'll use gnuplot for plotting, from the command line in Cygwin or Linux and also as a standalone program on Windows.
• The GSL ("Gnu Scientific Library") is a free numerical library.
• Python is available at the command line in Cygwin or Linux, and there is a stand-alone version on Windows.
• MATLAB and Mathematica are available on all platforms for registered OSU students.

Instructor:

Prof. Richard Furnstahl
office: M2048 PRB
email: furnstahl.1@osu.edu or furnstah@mps.ohio-state.edu
phone: 292-4830 (office) or 847-4026 (home)

1094 Consultant:

Dr. Leslie Schradin
office: TBA
email: schradin@mps.ohio-state.edu
phone: TBA

Computer Consultant:

Ms. Terry Bradley
office: 1199 PRB
email: bradleyt@mps.ohio-state.edu
phone: 292-8598 (PRB office) or 292-4269 (Stillman Hall)

Schedule:

Class meets MW from 2:30pm to 4:30pm or 3:30pm to 5:30pm in Smith 1094 (come either time). Each period will primarily be a hands-on lab session (after a short lecture/question part).

Office Hours:

By appointment (asking in class is easiest) and . . .
[to be announced] (Furnstahl)
[to be announced] (Schradin)

Grading:

In-class worksheets [30%]

Assigned homework [40%]

Project [30%]

Web Pages:

This info: http://www.physics.ohio-state.edu/~ntg/780/compphys_info.php

Course home page: http://www.physics.ohio-state.edu/~ntg/780/compphys.php