Adaptation of legacy codes to context-aware composition using aspect-oriented programming

摘要: The context-aware composition approach (CAC) has shown to improve the performance of object-oriented applications on modern multi-core hardware by selecting between different (sequential and parallel) component variants in (call hardware) contexts. However, introducing CAC legacy can be time-consuming requires quite some effort for changing adapting existing code. We observe that CAC-concerns, like offline variant profiling runtime selection champion variant, separated from application suggest separating reusing these concerns when applications. For automating this process, we propose an based Aspect-Oriented Programming (AOP) Reflective Programming. It shows manual adaptation more programming than AOP-based approach; almost three times our experiments. Moreover, speeds up execution time code, experiments factors 2.3 3.4 machines with two eight cores, respectively. AOP only introduces a small overhead compared manually optimized approach. For problems, is about 2-9% approach. These results effectively adapt which makes them running efficiently even machines.