renderingpipelinen

The rendering pipeline, also known as the graphics pipeline, is the sequence of processing steps used to convert a three-dimensional scene into a two-dimensional image on a display. It encompasses geometry processing, shading, rasterization, and final pixel output, and is largely executed on the GPU under orchestration by the CPU and drivers.

In real-time graphics, the pipeline proceeds through stages such as the input assembler, vertex shader, optional tessellation and geometry shaders, a rasterizer, and the fragment shader. After shading, depth and stencil tests, blending, and write to the framebuffer complete the core path. Additional passes may apply post-processing effects.

Historically, hardware started with fixed-function pipelines and later adopted programmable stages, enabling complex visual effects. Modern pipelines rely on shaders written in languages such as GLSL, HLSL, or SPIR-V, and APIs that drive programmable stages. The concept of the pipeline also extends to offline rendering, where path tracing and other techniques fit into a broader sequence of data transformation and shading.

Variants include tile-based renderers for mobile platforms, compute-based pipelines, and ray tracing pipelines that integrate ray tracing with rasterization. Contemporary APIs like DirectX 12, Vulkan, and Metal expose pipeline state objects to configure shaders, rasterization behavior, depth and stencil operations, and blending.

In practice, rendering often uses multi-pass or deferred shading, where geometry is drawn in one or more passes and lighting is computed in screen space, followed by post-processing effects such as tone mapping, bloom, and anti-aliasing. The rendering pipeline remains a foundational concept across graphics systems, informing both hardware design and software architecture.