Authors:
René Heinzl
;
Philipp Schwaha
and
Siegfried Selberherr
Affiliation:
Institute for Microelectronics, TU Wien, Austria
Keyword(s):
Scientific computing, high performance computing, concept based programming, multi paradigms, partial differential equations.
Related
Ontology
Subjects/Areas/Topics:
Applications and Software Development
;
Component-Based Software Engineering
;
Embedded Communications Systems
;
Model-Driven Software Development
;
Software Architectures
;
Software Engineering
;
Telecommunications
Abstract:
During the last decades various high-performance libraries were developed written in fairly low level languages, like FORTRAN, carefully specializing codes to achieve the best performance. However, the objective to achieve reusable components has been regularly eluded by the software community ever since. The fundamental goal of our approach is to create a high-performance mathematical framework with reusable domain-specific abstractions which are close to the mathematical notations to describe many problems in scientific computing. Interoperability driven by strong theoretical derivations of mathematical concepts is another important goal of our approach.