Mesh-oriented 3d Graphics Architecture

Although triangle meshes are used pervasively in 3D graphics applications and there exist highly eecient mesh representations, almost all existing 3D graphics processors are based on the assumption that individual triangles are processed completely independently of one another. Consequently, none of these graphics processors are able to exploit triangle mesh's vertex/edge sharing property. This paper describes a mesh-oriented 3D graphics architecture called Heresy, which treats meshes as rst-class objects , and signiicantly reduces the computation and communication costs of rendering triangle meshes by reusing intermediate computation results and eliminating data redundancy. The central architectural feature of Heresy is a highly eecient triangle mesh representation based on breadth-rst mesh traversal, which is applied throughout the entire 3D graphics pipeline, from geometric transformation, clipping, to rasteriza-tion, and thus enables aggressive exploitation of vertex/edge sharing to minimize both transformation/shading and rasterization cost. Results of a detailed simulation study based on a set of sophisticated 3D models demonstrate the superiority of Heresy over traditional triangle-based architectures and show that on the average the geometric transformation/shading cost, the communications overhead between CPU and the ras-terization hardware, and the rasterization delay are reduced by 82%, 74% and 37%, respectively. Moreover, owing to a window-based prefetching technique, the on-chip storage requirement for Heresy is shown to remain small (about 128 vertices wide) and largely independent of the size of the triangle mesh.