Quality of Numerical Software: Assessment and enhancementRonald F. Boisvert Numerical software is central to our computerized society. It is used to control aeroplanes and bridges, operate manufacturing lines, control power plants and refineries, and analyse financial markets. Such software must be accurate, reliable, robust, efficient, easy to use, maintainable and adaptable. Quality assessment and control of numerical software is still not well understood. Although measurement is a key element, it remains difficult to assess many components of software quality and to evaluate the trade-offs between them. Fortunately, as numerical software is built upon a long established foundation of mathematical and computational knowledge, there is great potential for dramatic breakthroughs. This volume will address enabling techniques and tools such as benchmarks, testing methodologies, quality standards, metrics, and accuracy control mechanisms, and their application to software for differential equations, linear algebra, data analysis, as well as the evaluation of integrals, derivatives and elementary and special functions. |
Contents
If software quality is a perception how do we measure | 32 |
A functional approach to software reliability modeling | 61 |
Quality of service and scientific workflows | 77 |
Improving the quality of software quality determination processes | 90 |
PART TWO Testing and Evaluation Methodology | 107 |
a web resource for test matrix collections | 125 |
A methodology for testing classes of approximation | 138 |
Evaluation of minimization software based on performance profile | 152 |
Why we couldnt use numerical libraries for PETSC | 249 |
Networkbased scientific computation via Inferno | 267 |
The XSC tools for extended scientific computing | 280 |
Is nonnormality a serious computational difficulty in practice? | 300 |
Reliability of local error control algorithms for initial value ordinary | 315 |
Software testing and evaluation in largescale scientific applications | 330 |
Some fundamental limitations of mathematical software revealed by | 345 |
Development of efficient general purpose Monte Carlo codes used | 349 |
A proposed software test service for special functions | 167 |
Two approaches to exception handling in Fortran | 210 |
Developing ODE software in new computing environments | 224 |
PART FIVE The Conference | 374 |
Other editions - View all
Quality of Numerical Software: Assessment and Enhancement Ronald F. Boisvert No preview available - 1997 |
Common terms and phrases
accuracy algorithm analysis application approach arguments arithmetic assessment automatic differentiation backward error Bischof BLAS compiler complex components Computer Science condition number Croz database defined described differential equations distribution domain Dongarra eigenvalue eigenvector error bounds estimate evaluation example execution factor failure Figure floating point format Fortran 77 Fortran 90 function grid Hammarling Harwell-Boeing Higham Hilbert matrix IEEE implementation input integration interface iterative KPA's Laboratory language LAPACK linear algebra linear system LINPACK Math mathematical software MATLAB Matrix Market measure method nonnormality Numerical Algorithms Group numerical software operations package parallel computer parameters performance precision problem procedure processor quality determination process reference data sets reference results reliability requirements routines ScaLAPACK scientific computing software development software processes software product software quality software testing solution solvers solving sparse matrix specific standard structure subroutine Technology test software user's variables vector