Reconfigurable Work Farms on a Massively Parallel Processor Array

A massively parallel processing array platform for reconfigurable computing is based on a structural object programming model. Objects are software programs running concurrently on hundreds of 32-bit RISC processors and memories. They exchange data and control through a structure of self-synchronizing channels. An IDE compiles source code and block diagrams into a configuration file in less than one minute. A common application design pattern on this platform, called a work farm, is a parallel set of worker objects, with one input and one output stream. Statically configured work farms with homogeneous and heterogeneous sets of workers have been used in video compression and decompression, network processing, and graphics applications. This work extends the programming model into dynamic runtime self-reconfiguration. First a general technique for dynamic objects that retain a static internal structure is developed, then it is extended to dynamic objects with dynamic structures. Reconfiguration speed and file size is orders of magnitude better than FPGA reconfiguration.