Refactoring sequential embedded software for concurrent execution using design patterns

Multicore architectures are becoming available in embedded systems. However, parallelizing sequential software is a challenging task. A structured approach is needed to exploit parallel opportunities. Therefore we propose a topdown approach based on a layered model of parallel design patterns. As a proof of concept this approach has been applied on a number of algorithms including the Fast Fourier Transformation. The FFT algorithm has been implemented parallel on a Linux based embedded system while others were implemented in a freestanding environment.status: publishe