Unified shaders

Unified shaders

Unified shaders refer to a graphics processing architecture that combines vertex and pixel shaders into a single stream processor, allowing for simultaneous processing and avoiding unbalanced loading. This architecture is supported by both Nvidia and AMD GPUs, which have become more flexible and can be reconfigured to support different shader stages in the graphical rendering pipeline. Shader Model 4 also features a unified shader architecture, in which all programmable stages of the pipeline share cores and resources, including graphics memory.From: Machinery, Materials Science and Engineering Applications [2017], Speeding Up Monte Carlo Computations by Parallel Processing Using a GPU for Uncertainty Evaluation in accordance with GUM Supplement 2 [2018], 3D Graphics for Game Programming [2019]

Exploitation and realization of 3D virtual scene based on DirectX 11 technology

View Chapter

Purchase Book

Published in Fei Lei, Qiang Xu, Guangde Zhang, Machinery, Materials Science and Engineering Applications, 2017

Qiong Yuan, Hui Zhang

As the latest multimedia programming interface of Microsoft, DirectX 11 [1] has been markedly reinforced on a basis of DirectX 10, which is a thorough reformation by contrast with DirectX 9 used in China. It is provided with unified shader architecture, while lots of graphics hardware and APIs use separated shader architecture, which means vertex shader and pixel shader are separated. The unified shader architecture in DirectX 11 combines vertex shader and pixel shader to achieve stream processor which can simultaneously process vertex shader and pixel shader and avoid unbalanced loading.

Speeding Up Monte Carlo Computations by Parallel Processing Using a GPU for Uncertainty Evaluation in accordance with GUM Supplement 2

View Article

Journal Information

Published in NCSLI Measure, 2018

C. M. Tsui, Aaron Y. K. Yan, H. W. Lai

GPUs are designed to render 3D computer graphics in real time. They have highly parallel architecture optimized for manipulating pixels in frame buffers. In the early days, GPUs had limited programmability and were seldom used outside the graphics processing area. The situation started to change in 2007. The architectures of GPUs from two major suppliers, Nvidia and AMD, became more flexible so that the hardware allowed re-configuration to support different shader stages (A “shader” is computer jargon referring to a short program for handling graphics.) in the graphical rendering pipeline (the unified shader model). People began to realize that with this flexibility, it became feasible to deploy the highly parallel architecture of GPUs for general scientific computing applications.

Explore chapters and articles related to this topic

Exploitation and realization of 3D virtual scene based on DirectX 11 technology

Speeding Up Monte Carlo Computations by Parallel Processing Using a GPU for Uncertainty Evaluation in accordance with GUM Supplement 2