The createBindGroup() method of the GPUDevice interface creates a GPUBindGroup based on a GPUBindGroupLayout that defines a set of resources to be bound together in a group and how those resources are used in shader stages.
createBindGroup(descriptor)
A GPUBufferBinding object can contain the following properties:
buffer -
The GPUBuffer object you want to bind.
-
offset Optional
-
The offset, in bytes, from the beginning of the buffer to the beginning of the range exposed to the shader by the buffer binding. If omitted, offset defaults to 0.
-
size Optional
-
The size, in bytes, of the buffer binding. If omitted, size will be the range starting at offset and ending at the end of the buffer. If both offset and size are omitted, the entire buffer is exposed to the shader.
A GPUBindGroup object instance.
The following criteria must be met when calling createBindGroup(), otherwise a GPUValidationError is generated and an invalid GPUBindGroup object is returned:
- The number of entries in the
layout GPUBindGroupLayout equals the number of entry objects in entries. - For each entry in the
layout GPUBindGroupLayout, the corresponding entry object in entries binds the correct resource type. For example, a buffer resource layout object has a GPUBufferBinding object specified in the corresponding binding. - If the resource layout object is a
buffer: - The corresponding bound
GPUBuffer: - Has its bound part (as specified by
offset and size) contained inside it completely, with a non-zero size. - Has a size bigger than the
buffer resource layout's minBindingSize.
- If the resource layout object
type is "uniform": - The bound
GPUBuffer has a usage that includes GPUBufferUsage.UNIFORM. - The effective size of the bound buffer segment is less than or equal to the
GPUDevice's maxUniformBufferBindingSize limit. - The specified
GPUBufferBinding offset is a multiple of the GPUDevice's minUniformBufferOffsetAlignment limit.
- If the resource layout object
type is "storage" or "read-only-storage": - The bound
GPUBuffer has a usage that includes GPUBufferUsage.STORAGE. - The effective size of the bound buffer segment is less than or equal to the
GPUDevice's maxStorageBufferBindingSize limit. - The effective size of the bound buffer segment is a multiple of 4.
- The specified
GPUBufferBinding offset is a multiple of the GPUDevice's minStorageBufferOffsetAlignment limit.
- If the resource layout object is a
storageTexture, the corresponding bound GPUTextureView: - Has a
dimension equal to the resource layout object's viewDimension (see GPUTexture.createView() for more details of a texture view's settings). - Has a
format equal to the resource layout object's sampleType. - Has a
mipLevelCount equal to 1. - Is a view of a
GPUTexture with a usage that includes GPUTextureUsage.STORAGE_BINDING.
- If the resource layout object is a
texture, the corresponding bound GPUTextureView: - Has a
dimension equal to the resource layout object's viewDimension (see GPUTexture.createView() for more details of a texture view's settings). - Has a
format compatible with the resource layout object's sampleType. - Is a view of a
GPUTexture with a usage that includes GPUTextureUsage.TEXTURE_BINDING. - is a view of a
GPUTexture with a sampleCount greater than 1 if the resource layout object's multisampled property is true, or equal to 1 if it is false.
Our basic compute demo shows an example of creating a bind group layout and then using that as a template when creating a bind group.
const bindGroupLayout = device.createBindGroupLayout({
entries: [
{
binding: 0,
visibility: GPUShaderStage.COMPUTE,
buffer: {
type: "storage",
},
},
],
});
const bindGroup = device.createBindGroup({
layout: bindGroupLayout,
entries: [
{
binding: 0,
resource: {
buffer: output,
},
},
],
});