Panel discussions: "Open source software and tools for engineering parallel software"
Supporting developers in creating parallel applications, re-engineering code from sequential ones, or factoring code for concurrency, scalability and fault-tolerance is essential for achieving sustainable performance improvement for upcoming heterogeneous and massively parallel computing systems. Here, the power of open source in parallel software development process is a strong driving force for developing modern parallel code and forms a core of the parallel software eco-system. Tools such as interpreters, translators, compilers and runtimes for parallel and concurrent platforms are now mostly linked to open source and community based development. Also, open source approaches are very important to enable reproducibility in parallel systems research. In this panel, we demonstrate the potential of open source software and seek its further roles for constructing and engineering parallel software toward upcoming computer systems.