渲染流水线中高速共享存储器的实现研究

Implementation of High Speed Shared Memory in Rendering Pipeline

作为GPU(图形处理器)中的核心,渲染流水线的执行速度决定了整个GPU的性能,其中通过设计和实现高速的共享存储器来保存渲染流水线中的中间结果,可以有效提升整个渲染流水线的性能。同时可以实现对CS(computer shader)相关指令的支持,提升GPU的计算能力。

As the core of GPU(Graphic Processor Unit),the execution speed of rendering pipeline determines the performance of the whole GPU.The performance of the whole rendering pipeline can be effectively improved by designing and implementing high-speed shared memory to save intermediate results in the rendering pipeline.At the same time,it can support the instruction of CS(computer shader)and improve the computing power of GPU.