3 A parallel incremental rendering method

Abstract

In this Chapter we have shown how a processor array can efficiently evaluate bivariate linear functions and thus implement rendering operations. Cheap, incremental, computation is thus exploited in conjuction with parallelism on a parallel architecture which has come to be recognised as general purpose. Our algorithms map very naturally onto the processor array because they are based on the window which is the processor array's primitive graphical object. The evaluation method can be extended to non-linear functions at a cost which is dependent on the order of the function.