Future computing systems will require dedicated accelerators to achieve high-performance. The mini-symposium ParaFPGA explores parallel computing with FPGAs as an interesting avenue to reduce the gap between the architecture and the application. Topics discussed are the power of functional and dataflow languages, the performance of high-level synthesis tools, the automatic creation of hardware multi-cores using C-slow retiming, dynamic power management to control the energy consumption, real-time reconfiguration of streaming image processing filters and memory optimized event image segmentation.