high performance computing on graphics processing units: hgpu.org

This short documentation was written by me for the High Performance Computing Seminar in the winter semester 2010/11 of Prof. G. Haase at the University Graz Austria. In order to learn Sphinx and to make it possible for others to get a quick start with Python in mathematics and scientific computing I started to write this tech report. I started with Python last summer, after a short introduction to the Sage mathematics software. One could say it was love at first sight. I was implementing some test code with krylov methods in Matlab and Octave that time, and was annoyed by the lack of object oriented features like abstracting and capsuling (and no: I don’t count structs as objects!). I had the problem, that every time I implement a new numerical scheme I had to rewrite the code of my optimisations algorithms, or at least have to alter it, so that every time I need to retest the new implemtation, which costs time and nerves. And since I’m a lazy person I didn’t want to do that. I’m now implementing my thesis code in Python, and I also work as a Sage developer in my freetime, and try to help improving the numerics, optimisations and symbolics parts of Sage which are my personal research interests. The Python version used here is Python 2.6. since most of the current packages are not ported to Python 3. But I try to make them compatible with the new version so that this document don’t get outdated soon. This document is aimed towards mathematicians who want learn Python in a quick way to use it. I have to admit that I’m a beginner too, and would be happy about feedback on this document. I plan to extend it frequently with tricks I learned and collected on various newsgroups. I don’t know If this is a good guide, but I hope potential readers have the same fun reading, I had writing.