| 摘要 |
A micro-vector processor has an input and output section, a supervisor section, a vector unit section, and a memory common to all of the sections and accessible by each. The signals to be processed are supplied by the input section in digital form to the data memory unit. The supervisor section is programmable and acts as the overall micro vector processor control processor. The supervisor unit sets up, starts, and monitors the actual signal processor, the vector unit. The supervisor orders the vector unit to go to a reserve location in the data memory unit where it finds the beginning address of a command-parameter list (CPL) previously loaded into data memory by the supervisor unit, containing the name of, and the parameters for, the utility algorithm which the vector unit is to execute. The algorithms that the vector unit performs are contained in the form of micro instructions in a read-only memory (ROM) which controls the operation of the vector unit. The vector unit sends results to the data memory unit and informs the supervisor that its task has been completed. The results are sent from the data memory to the output section. The vector unit, once started by the supervisor unit, is completely autonomous from the supervisor unit. The supervisor unit may control a plurality of vector units by simply invoking each one which then runs autonomously. Also, the entire micro-vector processor may be connected in an array with other micro-vector processors to form a distributed signal processing system. |
