Cluster Culling Shaders (CCS) are similar to the existing compute shaders. Their main purpose is to provide an execution environment in order to perform coarse-level geometry culling and LOD selection more efficiently on the GPU.

The traditional 2-pass GPU culling solution using a compute shader sometimes needs a pipeline barrier between compute and graphics pipeline to optimize performance. An additional compaction process may also be required. This extension addresses these shortcomings, allowing compute shaders to directly emit visible clusters to the following graphics pipeline.

A set of new built-in output variables are used to express a visible cluster. In addition, a new built-in function is used to emit these variables from CCS to the IA stage. The IA stage can use these variables to fetches vertices of a visible cluster and drive vertex shaders to shading these vertices.

Note that CCS do not work with geometry or tessellation shaders, but both IA and vertex shaders are preserved. Vertex shaders are still used for vertex position shading, instead of directly outputting transformed vertices from the compute shader. This makes CCS more suitable for mobile GPUs.

Sample code

Example of cluster culling in a GLSL shader

#extension GL_HUAWEI_cluster_culling_shader: enable

#define GPU_WARP_SIZE                   32
#define GPU_GROUP_SIZE                  GPU_WARP_SIZE

#define GPU_CLUSTER_PER_INVOCATION      1
#define GPU_CLUSTER_PER_WORKGROUP       (GPU_GROUP_SIZE * GPU_CLUSTER_PER_INVOCATION)

// Number of threads per workgroup
// - 1D only
// - warpsize = 32
layout(local_size_x=GPU_GROUP_SIZE, local_size_y=1, local_size_z=1) in;


#define GPU_CLUSTER_DESCRIPTOR_BINDING      0
#define GPU_DRAW_BUFFER_BINDING             1
#define GPU_INSTANCE_DESCRIPTOR_BINDING     2

const float pi_half = 1.570795;
uint instance_id;

struct BoundingSphere
{
  vec3 center;
  float radius;
};

struct BoundingCone
{
  vec3 normal;
  float angle;
};

struct ClusterDescriptor
{
  BoundingSphere sphere;
  BoundingCone cone;
  uint instance_idx;
};

struct InstanceData
{
  mat4 mvp_matrix;                      // mvp matrix.
  vec4 frustum_planes[6];               // six frustum planes
  mat4 model_matrix_transpose_inverse;  // inverse transpose of model matrix.
  vec3 view_origin;                     // view original
};

struct InstanceDescriptor
{
  uint begin;
  uint end;
  uint cluster_count;
  uint debug;
  BoundingSphere sphere;
  InstanceData instance_data;
};

struct DrawElementsCommand{
  uint indexcount;
  uint instanceCount;
  uint firstIndex;
  int  vertexoffset;
  uint firstInstance;
  uint cluster_id;
};

// indexed mode
out gl_PerClusterHUAWEI{
  uint gl_IndexCountHUAWEI;
  uint gl_InstanceCountHUAWEI;
  uint gl_FirstIndexHUAWEI;
  int  gl_VertexOffsetHUAWEI;
  uint gl_FirstInstanceHUAWEI;
  uint gl_ClusterIDHUAWEI;
};


layout(binding = GPU_CLUSTER_DESCRIPTOR_BINDING, std430) readonly buffer cluster_descriptor_ssbo
{
        ClusterDescriptor cluster_descriptors[];
};


layout(binding = GPU_DRAW_BUFFER_BINDING, std430) buffer draw_indirect_ssbo
{
        DrawElementsCommand draw_commands[];
};

layout(binding = GPU_INSTANCE_DESCRIPTOR_BINDING, std430) buffer instance_descriptor_ssbo
{
        InstanceDescriptor instance_descriptors[];
};


bool isFrontFaceVisible( vec3 sphere_center, float sphere_radius, vec3 cone_normal, float cone_angle )
{
  vec3 sphere_center_dir = normalize(sphere_center -
                           instance_descriptors[instance_id].instance_data.view_origin);

  float sin_cone_angle = sin(min(cone_angle, pi_half));
  return dot(cone_normal, sphere_center_dir) < sin_cone_angle;
}

bool isSphereOutsideFrustum( vec3 sphere_center, float sphere_radius )
{
  bool isInside = false;

  for(int i = 0; i < 6; i++)
  {
      isInside = isInside ||
      (dot(instance_descriptors[instance_id].instance_data.frustum_planes[i].xyz,
      sphere_center) + instance_descriptors[instance_id].instance_data.frustum_planes[i].w <
      sphere_radius);
  }
  return isInside;
}


void main()
{
    uint cluster_id = gl_GlobalInvocationID.x;
    ClusterDescriptor desc = cluster_descriptors[cluster_id];

    // get instance description
    instance_id = desc.instance_idx;
    InstanceDescriptor inst_desc = instance_descriptors[instance_id];

    //instance based culling
    bool instance_render = !isSphereOutsideFrustum(inst_desc.sphere.center, inst_desc.sphere.radius);

    if( instance_render)
    {
        // cluster based culling
        bool render = (!isSphereOutsideFrustum(desc.sphere.center,
        desc.sphere.radius) && isFrontFaceVisible(desc.sphere.center, desc.sphere.radius, desc.cone.norm
        al, desc.cone.angle));

        if (render)
        {
            // this cluster passed coarse-level culling, update built-in output variable.
            // in case of indexed mode:
            gl_IndexCountHUAWEI     = draw_commands[cluster_id].indexcount;
            gl_InstanceCountHUAWEI  = draw_commands[cluster_id].instanceCount;
            gl_FirstIndexHUAWEI     = draw_commands[cluster_id].firstIndex;
            gl_VertexOffsetHUAWEI   = draw_commands[cluster_id].vertexoffset;
            gl_FirstInstanceHUAWEI  = draw_commands[cluster_id].firstInstance;
            gl_ClusterIDHUAWEI      = draw_commands[cluster_id].cluster_id;

            // emit built-in output variables as a drawing command to subsequent
            // rendering pipeline.
            dispatchClusterHUAWEI();
        }
    }
}

Example of graphics pipeline creation with cluster culling shader

// create a cluster culling shader stage info structure.
VkPipelineShaderStageCreateInfo ccsStageInfo{};
ccsStageInfo.sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO;
ccsStageInfo.stage = VK_SHADER_STAGE_CLUSTER_CULLING_BIT_HUAWEI;
ccsStageInfo.module = clustercullingshaderModule;
ccsStageInfo.pName =  "main";

// pipeline shader stage creation
VkPipelineShaderStageCreateInfo shaderStages[] = { ccsStageInfo, vertexShaderStageInfo, fragmentShaderStageInfo };

// create graphics pipeline
VkGraphicsPipelineCreateInfo pipelineInfo{};
pipelineInfo.sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO;
pipelineInfo.stageCount = 3;
pipelineInfo.pStage = shaderStages;
pipelineInfo.pVertexInputState = &vertexInputInfo;
// ...
VkPipeline graphicsPipeline;
VkCreateGraphicsPipelines(device, VK_NULL_HANDLE, 1, &pipelineInfo, nullptr, &graphicsPipeline);

Example of launching the execution of cluster culling shader

vkCmdBindPipeline(commandBuffer, VK_PIPELINE_BIND_POINT_GRAPHICS, graphicsPipeline);
vkCmdDrawClusterHUAWEI(commandBuffer, groupCountX, 1, 1);
vkCmdEndRenderPass(commandBuffer);

VK_HUAWEI_cluster_culling_shader

Name String

VK_HUAWEI_cluster_culling_shader

Extension Type

Device extension

Registered Extension Number

405

Revision

2

Extension and Version Dependencies

VK_KHR_get_physical_device_properties2

Contact

• Yuchang Wang richard_Wang2

Extension Proposal

VK_HUAWEI_cluster_culling_shader

Other Extension Metadata

Last Modified Date

2022-11-17

Interactions and External Dependencies

• This extension requires SPV_HUAWEI_cluster_culling_shader.
• This extension provides API support for GL_HUAWEI_cluster_culling_shader.

Contributors

• Yuchang Wang, Huawei
• Juntao Li, Huawei
• Pan Gao, Huawei
• Jie Cao, Huawei
• Yunjin Zhang, Huawei
• Shujie Zhou, Huawei
• Chaojun Wang, Huawei
• Jiajun Hu, Huawei
• Cong Zhang, Huawei