Component Software: Beyond Object-Oriented Programming
Component Software: Beyond Object-Oriented Programming
Iterative Methods for Sparse Linear Systems
Iterative Methods for Sparse Linear Systems
Component-based iterative methods for sparse linear systems: Research Articles
Concurrency and Computation: Practice & Experience - Component and Framework Technology in High-Performance and Scientific Computing
Bocca: a development environment for HPC components
Proceedings of the 2007 symposium on Component and framework technology in high-performance and scientific computing
Usability levels for sparse linear algebra components
Concurrency and Computation: Practice & Experience
| Hi-index | 0.00 |
Component interfaces for sparse linear system solvers facilitate their integration into scientific applications. Due to a great variety of sparse matrix representations and characteristics, the solver choice may be complicated. Furthermore, once chosen, the solver often needs tuning to a given application. Thus, the tasks of solver switching (on the high level of abstraction) and solver tuning may be combined calling for hierarchical linear algebra interfaces. In other words, the user may access the solver either as a whole package or as a set of tunable components. In our previous work, we proposed common interfaces for these two types of accesses. These interfaces were implemented using Common Component Architecture (CCA) specifications. In the current work, we provide and implement a hierarchical design for parallel linear system components and show the benefits of accessing sparse linear algebra functionality via integrated high- and medium-level interfaces.