Design and implementation of fully programmable portable 3D graphics processor with mobile unified shader

Abstract

The real time 3D graphics becomes the most attractive application for mobile terminals, in which the battery lifetime and small computing power. Recently, incorporation with 3D navigation services and wireless 3D games promotes to realize programmable 3D graphics pipeline in mobile devices. In addition, since users watch graphics images on a small screen very closely to their eyes, recent mobile 3D graphics are introducing the programmability in both hardware and software for more advanced functionality while achieving low power consumption. In this research, I designed and implemented a low power fully programmable 3D graphics processor with mobile unified shader for mobile applications. It can provide per-pixel operations and per-pixel operations in a single hardware and thus, it generates realistic 3D graphics effects with low power consumption and small area. The proposed 3D graphics processor has four major features: mobile unified shader, full hardware accelerations with stream processing, logarithmic datapath, and unified SIMD datapath. The 3D graphics processor consists of mobile unified shader, vertex generator, fragment generator, pixel generator, and graphics cache system. And it was embedded into ARM9 based mobile multimedia SoC and the SoC was fabricated using 0.13μm 1Poly 7Metal CMOS logic process and it consumed $36 mm^2$. The 3D graphics processors contains 1.3M logic gates and it consumes 3.3mm $\times$ 3.0mm. With the mobile unified shader, the developed SoC can provide fully programmable 3D graphics pipeline, per-vertex operations and per-pixel operations, and thus, it can completely generate realistic 3D graphics effects such as environmental mapping or per-pixel lighting. The proposed mobile unified shader reduces silicon area of the 3D graphics processor by 35% with the help of the unified SIMD datapaths, logarithmic datapath. The logarithmic datapaths, low power register file, and micro-level (instruction-level and pixel-level) clo...