Programming Techniques for Scientific Simulations II

This course covers advanced general and C++ programming techniques relevant for scientific simulations. The course will cover, in particular:

• generic algorithm and library design
• exception safety
• smart pointers and safe memory handling
• polymorphism at compile time, at run time and hybrid designs
• mixed language programs, in particular C++, C, Fortran and Python, and the Boost.Python library
• template meta programming and relevant libraries
• C++ libraries for parallel programming on distributed and shared memory machines
• useful C++ libraries from Boost and other sources

Credit requirement

The class consists of lectures and exercises. You are expected to attend the classes and work on the exercises. Some exercises you can work on in class and directly show to the assistant. Others need to be done or finished as homework and sent to the assistant by e-mail.

Exercises

Exercise sheets and code fragments can be obtained from the subversion repository svn+ssh://username@plimpy.ethz.ch/home/gukel/SVNREPS/pt2. You need an account on the D-PHYS machines and an installation of the subversion version control system on your local machine.
In case plimpy is not accessible, you can replace it with any other D-PHYS Linux host, e.g. plumpy, plompy or plempy.

Useful references for this course

• How to get an account on the D-PHYS machines
• Subversion Version Control System
• Boost C++ Libraries