It has been shown that the existing von-Neumann design machines fail to meet the requirements of the real-time scientific and the 5th generation applications. A computer organization, such as data driven architecture, which embeds parallelism in its definition and underlying design is needed to satisfy such requirements. In this project we will introduce a new data driven multiprocessor which: (i) eliminates some of the problems of the existing data driven systems; (ii) takes advantage of state-of-the-art technology; and, (iii) achieves parallelism at the data, instruction, and program block levels. The system consists of n processing modules, a host module, and a data structure module, connected via a double star network with the host and the data structure modules in the center of the stars. Processing modules are capable of the simultaneous execution of independent program blocks in data driven fashion. The host module is responsible for memory management tasks, while the data structure module facilitates the manipulation of the data structures.