In response to the increasing complexity of realtime embedded software, driven by the need for intelligent computation in constrained environments, multicore architectures have emerged as a promising solution. The challenge lies in to choose an effective mapping of the real-time tasks to the various computational resources of these embedded boards, while ensuring real-time constraint satisfaction. To address this problem, our approach rests on two pillars. The first is a domainspecific language designed to capture hardware and software characteristics, constraints, and criteria in a clear and unambiguous manner. The second is a solver method based on Satisfiability Modulo Theories solver augmented with a lazy theory to handle real-time aspects. This method allows us to synthesize mappings that respect temporal constraints and optimize specific criteria, such as the power consumption of the embedded board.