The SAGA project will demonstrate that algebraically oriented programming methodologies provide means to reduce the time and effort needed for mathematical modelling, software development and porting to HPC architectures. This will be achieved through an increased ability to prototype and experiment with solution strategies and a more efficient, reliable and mature design of programs and software libraries. To do this well, a precise identification of the problems and conditions in which these techniques may be beneficially applied is needed.
Work will focus on three topics:
- Quick prototyping of mathematical models: SAGA will provide a tool based on algebraic model theory to translate computational models formulated as recurrences into non-recursive, efficient programs.
- Letting users experiment with different solution strategies and HPC architectures without the need for extensive reprogramming: SAGA will provide a methodology for programming co-ordinate free PDE solvers using a software library of interchangeable modules implemented on a range of HPC architectures.
- Allowing programmers to focus more on the problem to be solved and less on tricks to increase run-time efficiency: SAGA will provide a rule based code efficiency improvement tool that will transform simple, straight forward code into efficient code.
Computational modelling using partial differential equations (PDEs) is becoming an important tool for industry. Currently this is hampered by the time and effort needed to develop good computational models, and the time and cost needed to port such models onto a high performance computer (HPC).