A transputer architecture for parallel processing of polygonal regions

The growing interest in computer graphics has to be faced with the great complexity of many practical applications, as for example digital mapping. This paper discusses a solution in the field of parallel management of polygonal regions. The proposal is designed according with an integrated approach towards the definition of a data representation scheme suitable to be processed by parallel algorithms and its implementation on a parallel architecture programmed by a concurrent language. The proposed solution is based on DPCS, a distributed version of a data representation scheme called PCS. Data processing operates on partitioned data structures, and a transformation method is defined by a parallel algorithm to convert data represented as vectorial information to and from DPCS. The DPCS system is implemented on a Transputer ring, by now composed of five Transputers, but easily extendible. The processing and communicating facilities of the Transputer permit to exploit the maximum degree of parallelism, not only when processing the partitioned data, but also during the inherently sequential operations of data conversion.