|Software Infrastructure for Computational Science||
Development of new mathematical algorithms and software design
technologies which enable the development and maintenance of
applications on high performance computing architectures.
Emphasis on core areas such as linear algebra, and on mathematical domains of particular concern to NIST programs where significant gaps in software infrastructure exist.
|Testing and Evaluation Infrastucture for Mathematical Software||
Development of techniques and tools which support unbiased testing and
evaluation of mathematical algorithms and software by researchers, developers
Emphasis on development and compilation of reference data, reference algorithms and related evaluation methodology for core areas such as linear algebra and special function evaluation, and their widespread dissemination using modern network technologies.