Mesh Shaders and the History of Computer Graphics
From CPU and DSP software in the 70s, to fixed function dedicated hardware in the early 80s and then to fully integrated GPUs in the late 90s, transform and lighting functions have evolved from assembly language tasks to higher-level abstractions known as shaders. But the concept of a linear pipeline stayed with us from the 70s to 2021. With mesh shaders we have moved to a full compute mode for shading which includes transformations, tessellations, textures, and baked in rendering. Mesh shaders expose the full power of generalized GPU compute capability to the geometry processing pipeline, unlocking a new level of flexibility and performance.
The main goal of the Mesh shader is to increase the flexibility and performance of the geometry pipeline. Mesh shaders subsume most aspects of vertex and geometry shaders into one shader stage by processing batches of vertices and primitives before the rasterizer. They are additionally capable of amplifying and culling geometry.
Mesh shaders also enhance performance by allowing geometry to be pre-culled without outputting new index buffers to memory, whereas currently some geometry is culled by fixed function hardware.
From the hardware perspective the goal is to remove the need for the index processing part of the input-assembler (IA) and allow GPUs to be more parallel.
Discover Tech News: https://www.computer.org/publications...