Key points are not available for this paper at this time.
Described is a procedure for executing a visible surface algorithm in a new multi-microprocessor system which utilizes distributed image and depth (“Z”) buffers. It is shown that despite image distribution over a large number of processing and memory units, object coherence can still be maintained and used to reduce the number of calculations needed to generate a continuous-tone visible surface image.
Henry Fuchs (Sat,) studied this question.