Synergies in scientific computing by combining multi-paradigmatic languages for high-performance applications

Abstract

The challenging art of multi-paradigmatic application development, which only few languages currently support, greatly aids the development of highly efficient and reusable software components. A link of two such languages, namely Python and C++, is presented. Thereby data structures and algorithms realised in C++ using features such as compile-time meta-programming are made available to the run-time environment of Python. Several generic components and modules for application design in the area of scientific computing are presented. Compile times and run-times are discussed to show the advantages of the proposed combination of both languages.

Keywords:

• high performance programming
• scientific computing
• multi-paradigmatic programming
• Python
• C++

Acknowledgement

This work has been supported by the Austrian Science Fund FWF, project P19532-N13.