Scientific Software is a key component in developing effective instruments for Computational Science. Prototyping and developing scientific codes in terms of reliable, efficient and portable building blocks allow users to reduce the time-to-solution of a computational problem and to simplify the inclusion and comparisons of new physical/mathematical models and solvers. The history of open-source high-perfomance scientific libraries and frameworks, such as ScaLapack, FFTW, PETSc, Trilinos, ATLAS, only to cite some of the most widely used, is a history of success. Some of the mentioned packages represent both "de facto" standard platforms for scientific code development and benchmarks for new proposals of hardware/software architectures devoted to high-performance computing. On the other hand, in the new era of highly parallel heterogeneous architectures integrating homogeneous multicore CPUs and GPU type components, it is necessary to rethink standard software platforms.
Designing and implementing high-quality, reusable, extensible and portable scientific software are challenging tasks, requiring inputs and skills from different areas of Mathematics and Computer Science. Many issues have to be considered, such as targeting theoretical efficiency when designing new algorithms, being at the same time aware of architectural limitations; analyzing performance of such algorithms on emerging computers by realistic performance models as well as by practical implementation; using advanced programming tools for simplicity of usage and portability of the resulting software.
The focus of this workshop is on recent advances in algorithms and programming tools development for next-generation high-performance scientific software as enabling technologies for new insights into Computational Science.
The workshop is intended to bring together applied mathematicians, computer scientists and computational scientists from different areas, in order to discuss recent challenges and results in modern technology issues for high-performance computing as well as in developing open-source high-quality software for Computational Science.