Software Practical
"Implementation of rigorous tests for a linear algebra package."
Description
As a part of my Ph.D. project, new algorithms for mixed--integer optimal control
have been implemented. These algorithms heavily rely on a newly developed software package for
linear algebra, implemented in C. This package provides vector and matrix operations, matrix
decompositions, sparse matrices in triplets format, indexed vectors and matrices, and couples
to F77 BLAS / C BLAS / GSL C BLAS / ATLAS as well as F77 LAPACK / C LAPACK.
Aim
The aim of this software practical is to implement an extensive set of rigorous tests for
the linear algebra software package. The tests shall cover all provided operations and verify
that they operate correctly. Special care shall be taken to cover border cases (such as
matrices or vectors with dimension zero), and to verify that errors are signalled correctly
(such as failure of positive definiteness during Cholesky decomposition).
Major milestones of the software practical include
- Installation of the software package on the machine used for carrying out the practical.
- Generating, reading, and understanding the existing doxygen documentation of the linear algebra package.
- Determining categories of linear algebra operations.
- Identifying tests, border cases, and error conditions for all operations.
- Implementation of the identified tests in C.
- Documentation of the implementation in doxygen format.
Focus
Software implementation, linear algebra.
Size of the project
This project is suited for one or two students (80 hours per person) as an advanced software practical.
The software practical can be carried out in the PC pool of IWR or using your own PC / notebook.
We will meet regularly (ideally weekly) to discuss your progress and define your next points of work.
The practical shall be completed by the start of the next semester.
Requirements
- Firm knowledge of C is required.
- Knowledge of linear algebra and numerics (lectures 'Lineare Algebra 1', 'Einführung in die Numerische Mathematik') is required.
Contact
Dipl.-Math. Christian Kirches
Interdisciplinary Center for Scientific Computing (IWR)
Im Neuenheimer Feld 368
Universität Heidelberg
e-mail: christian.kirches@iwr.uni-heidelberg.de
Room : INF 368 (IWR), R 414
back
|