An effective rasterization architecture for mobile vector graphics processors

Abstract

This paper proposed a novel index board rasterization architecture which reduces mathematical calculations and memory traffic for vector graphics. The proposed architecture uses the cell based method which has advantages in computational complexity, and generates the active span by referring to only valid cells and placing them in scanline order with two internal SRAMs. The proposed architecture reduces the amount of calculation by an average of 59.4% and also the external memory traffic by an average of 30.0% compared to the traditional architecture.