The increasing complexity of signal processing algorithms has lead to the need of developing the algorithms specifications using generic software implementations that become in practice the reference implementation. This fact can be particularly observed in the field of video and multimedia processing where reference software is the main normative reference. Adapting the algorithms specified by such software models into architectures composed by processors and dedicated HW elements becomes a very resource consuming task for the complexity of the models and for the large choice of possible partitioning options. This paper describes a new platform aiming at supporting the adaptation of algorithms specified by generic non optimized software specifications into mixed SW and HW implementations. The platform is supported by profiling capabilities specifically developed to study data transfers between the SW and the HW modules. Such profiling and optimization capabilities can be used to achieve different objectives in the algorithm architecture adaptation process such as optimization of memory architectures or low power designs by the minimization of data transfers.