GPU 实例化

GPU instancing is a draw call optimization method that renders multiple copies of a mesh with the same material in a single draw call. Each copy of the mesh is called an instance. This is useful for drawing things that appear multiple times in a scene, for example, trees or bushes.

GPU instancing renders identical meshes in the same draw call. To add variation and reduce the appearance of repetition, each instance can have different properties, such as Color or Scale. Draw calls that render multiple instances appear in the Frame Debugger as Draw Mesh (instanced).

Requirements and compatibility

This section includes information about the platform, render pipeline, and SRP Batcher compatibility of GPU instancing.

Platform compatibility

GPU instancing is available on every platform except WebGL 1.0.

Render pipeline compatibility

注意：

1. Only if the shader isn’t compatible with the SRP Batcher.

SRP Batcher

GPU instancing isn’t compatible with the SRP Batcher. The SRP Batcher takes priority over GPU instancing. If a GameObject is compatible with the SRP Batcher, Unity uses the SRP Batcher to render it, not GPU instancing. For more information about optimization method priority, see Optimization priority.

If your project uses the SRP Batcher and you want to use GPU instancing for a GameObject, you can do one of the following:

• Use Graphics.DrawMeshInstanced. This API bypasses the use of GameObjects and uses the specified parameters to directly draw a mesh on screen.
• Manually remove SRP Batcher compatibility. For information on how to do this, see Intentionally removing compatibility.

Using GPU instancing

Unity uses GPU instancing for GameObjects that share the same mesh and material. To instance a mesh and material:

• The material’s shader must support GPU instancing. Unity’s Standard Shader supports GPU instancing, as do all surface shaders. To add GPU instancing support to any other shader, see [Creating shaders that support GPU instancing(gpu-instancing-shader).
• The mesh must be from one of the following sources:
  • A MeshRenderer component or a Graphics.DrawMesh call. Unity adds meshes from these sources to a list and attempts to instance them together. If the GameObject that a MeshRenderer component is attached to is SRP Batcher compatible, then Unity can’t instance that mesh. For more information, see SRP Batcher compatibility.
    Note: Unity doesn’t support GPU instancing for SkinnedMeshRenderers.
  • A Graphics.DrawMeshInstanced or Graphics.DrawMeshInstancedIndirect call. These methods instance the same mesh in multiple places using the same shader. Unity executes each call separately which means it can’t instance meshes from different calls together.

To use GPU instancing for a material, select the Enable GPU Instancing option in the Inspector.

Lighting

GPU instancing supports Unity’s Baked Global Illumination system. Unity Standard Shaders and surface shaders support GPU instancing and Unity’s Baked Global Illumination system by default.

Each GPU instance supports global illumination from one of the following sources:

• Any number of Light Probes.
• One lightmap.
  Note: An instance can use multiple atlas regions in the lightmap.
• One Light Probe Proxy Volume(LPPV) component.
  Note: You must bake the LPPV for the space volume that contains all the instances.

GPU instancing automatically works with:

• Dynamic Mesh Renderers affected by Light Probes.
• Static Mesh Renderers you bake to the same lightmap texture. A Mesh Renderer is static in this context if it includes Contribute GI in its Static Editor Flags.

To enable Light Probe rendering for Graphics.DrawMeshInstanced, set the LightProbeUsage parameter to CustomProvided and provide a MaterialPropertyBlock that includes the Probe data. For more information and code examples, see LightProbes.CalculateInterpolatedLightAndOcclusionProbes.

Alternatively, you can pass an LPPV component reference and LightProbeUsage.UseProxyVolume to Graphics.DrawMeshInstanced. When you do this, all instances sample the volume for the L0 and L1 bands of the Light Probe data. If you want to supplement L2 data and occlusion data, use a MaterialPropertyBlock. For more information, see Light Probes: Technical Information.

Performance implications

Meshes that have a low number of vertices can’t be processed efficiently using GPU instancing because the GPU can’t distribute the work in a way that fully uses the GPU’s resources. This processing inefficiency can have a detrimental effect on performance. The threshold at which inefficiencies begin depends on the GPU, but as a general rule, don’t use GPU instancing for meshes that have fewer than 256 vertices.

If you want to render a mesh with a low number of vertices many times, best practice is to create a single buffer that contains all the mesh information and use that to draw the meshes.

• 2017–10–24 页面已修订

• 在 5.6 版中增加了 Enable Instancing 复选框指南、DrawMeshInstancedIndirect 和 #pragma multi-compile

• 在 2017.3 版中增加了针对 GPU 实例化的着色器预热 NewIn20173

• 在 2018.1 版中增加了 GPU 实例化的全局光照 (GI) 支持 NewIn20181