Using Design Patterns in Scientific Applications
Аннотация:Revision with unchanged content. Design Patterns have been widely adopted for building flexible and extensible applications. However, this can come at a cost of reduced performance, which may not be acceptable for computationally intensive scientific applications. Therefore, the effects of design patterns on application performance are investigated through a detailed measurement and profiling of CompuCell3D. It is a software framework for three-dimensional modeling (3D) of morphogenesis, which is a stage in embryonic development where cells cluster into tissues and organs. By reverse engineering CompuCell3D subsystems with and without design patterns, their impact on application performance andmaintainability is evaluated. Finally, CompuCell3D functionality is extended to support irregular domains for biologically realistic simulations.