Unified Shader

As the manipulation of vertex and pixel becomes more complicated, today’s graphics processors are becoming more programmable like general CPUs. So, fixed pipelined GPU becomes shader processor with programmability. Furthermore, features of vertex/pixel shader in GPU are integrated in single programmable shader architecture, and nowadays this is called as a unified shader architecture. The latest graphics APIs such as Direct X 10 or Direct X 11 introduce the Geometry Shader for modification of geometry and the Compute Shader for general purpose computing, scientific calculation, financial computation, physical simulation, etc. The unified shader includes all the features of those shaders and it brings the new computation technology, GPGPU(General-Purpose computing on Graphics Processing Unit). We are focused on the efficient low-power unified shader architecture for mobile computing as well as the many-core unified shader architecture for high performance GPGPU and massively parallel computations.