During the last few years, the Radar Techniques Group at Lincoln Laboratory has been applying digital processing techniques to the problem of automatic detection of moving vehicles in the presence of ground and weather clutter. An outgrowth of this effort is the development of a real-time radar signal processor, the Parallel Microprogrammable Processor, or PMP. Conceptually the PMP consists of a single control unit and an array of identical processing modules. The control unit sequences through a program stored in its control memory, providing identical instructions to each processing module, so that all modules are performing the same operation in parallel, each on its own set of data. The talk will focus on the motivation for, and advantages of such a parallel architecture, as presently implemented with TTL medium-scale integrated circuits. Some examples of parallel computation will be illustrated as well as more general issues relating to programmability of the PMP. Much of the information in the talk will be based on experience with an operational prototype, which has a control unit and one processor module.