This course is organized by the Swedish e-Science Education program (SeSE).

The course consists of three parts - self reading, lectures and labs, and project preparation. The lectures and the labs will be given at the Department Information Technology, Uppsala University.

Preparatory work for the self study during Week 47:

As preparation for the course, we ask you to review the literature. Reading this material in full is much than a single week of work, but we will expect you to be somewhat familiar with what the intended scope is of the different technologies, based on our recommended sources and other material. We also ask you to brush up on your general Python and programming skills.

Course materials (will be updated currently during the course)

• Carl Nettelblad: Slides 1 .. Slides 2 .. Lab 1 .. Lab 2
• Ali Dorostkar: Slides and Lab materials
• Maya Neytcheva: Slides

Information on the project to be done in order to pass the course with full credit:

Detailed time schedule

Date	Topic(s)	Time	Location	Lecturer
Nov 23	General description of the course	9:15-9:30	6143	MN
	Introduction. High performance models and measures	9:30-12:00	6143	MN
	Models, measures, cont.	13:15-14:00	1113	MN
	Participants shortly describe their research project and the HPC-related issues	14:15-15:30	1113	MN
Nov 24	C++11, not C with classes anymore. Including templates, lambdas, smart pointers. Library use including the Standard Template Library, Boost, and interaction with Python.	9:15-12:00	6143	CN
	Computer lab and 'Hands-on' session: Create a C++ library called from Python using Cython and Boost Python.	13:15-17:00	1211	CN
Nov 25	Discussion of the experience from the last computer lab	9:15-9:30	2345	CN
	Design patterns and modular software - designing, building and maintaining. git and cmake	9:30-12:00	2345	CN
	Computer lab and 'Hands-on' session: Extending our Python library with source code management and a build system with external dependencies.	13:15-17:00	1112	CN
Nov 26	Discussion of the experience from the last computer lab	9:15-9:30	6111	CN
	An overoview of libraries for numerical simulations: dealii, Trilinos, Paralution, Armadillo, etc...	9:30-12:00	6111	AD
	Computer lab and 'Hands-on' session:	13:15-17:00	1312	AD
Nov 27	Parallel debuggers. Performance analysis tools: Allinea DDT, hands on	9:00-12:00	2345	MN/AD/Allinea
	Closing: details on the project work	13:30-15:00	1245	MN/AD/CN

Recommended sources:

1. Stroustrup, B. The C++ Programming Language (4th edition!), Addison-Wesley.
2. Gamma et al. Design Patterns, Elements of Reusable Object-oriented Software, Addison-Wesley
3. W. Bangerth, G. Kanschat, and R. Hartmann, deal.II differential equations analysis library, http://www.dealii.org.
4. M. A. Heroux and J. M. Willenbring, Trilinos Users Guide, http://trilinos.sandia.gov, Technical Report SAND2003-2952, Sandia National Laboratories, 2003. Matrix Methods in Data Mining and Pattern Recognition. SIAM, Philadelphia, PA, Philadelphia, PA, USA, 2007.
5. Armadillo C++ Linear Algebra Library, http://arma.sourceforge.net/
6. BOOST C++ Libraries, http://www.boost.org
7. Atlassian git tutorial, https://www.atlassian.com/git/tutorials/
8. cmake tutorial, https://cmake.org/cmake-tutorial/
9. Allinea DDT, http://www.allinea.com/products/ddt
10. TotalView, http://www.roguewave.com/products-services/totalview
11. D. Lurkarski Paralution project, http://www.paralution.com

Project

After the main lecture week, you are supposed to do a minor project based on what we have covered. Our main suggestion is to write a minor Python library for a performance-critical component of relevance in your scientific application. The library should be source-controlled, have a proper build environment, and ideally use modern scientific C++ libraries for parts of its operation.