Special session: Advanced tools for parallel and distributed programming

Compared to sequential programs, the execution behavior of parallel and distributed programs generally is much harder to predict during program development and can sometimes be surprising even for skilled programmers. Runtime tools, which monitor a program's execution and allow the programmer to systematically analyze its run-time behavior have become mature and useful instruments to address this problem. However, there is a high demand for still more advanced tools that amongst other things address user-friendliness by including an automatic analysis process or even an automatic feed-back, or support a more detailed analysis of an application's behavior, e.g. with respect to the memory hierarchy. This special session includes four research papers that present new ideas and advanced techniques, which are not yet found in established programming tools. Three of these contributions deal with performance analysis or prediction, indicating that on the one hand there is a high demand for sophisticated performance tools, on the other hand there are still many interesting research issues in this area.