GPUComputePassEncoder: setBindGroup() method

index

The index to set the bind group at. This matches the n index value of the corresponding @group(n) attribute in the shader code (GPUShaderModule) used in the related pipeline.

bindGroup

The GPUBindGroup to use for subsequent compute commands, or null, in which case any previously-set bind group in the given slot is unset.

dynamicOffsets Optional

A value specifying the offset, in bytes, for each entry in bindGroup with hasDynamicOffset: true set (i.e. in the descriptor of the GPUDevice.createBindGroupLayout() call that created the GPUBindGroupLayout object that the bindGroup is based on). This value can be:

• An array of numbers specifying the different offsets.
• A Uint32Array containing numbers specifying the offsets.

If a Uint32Array value is specified for dynamicOffsets, both of the following parameters are also required:

dynamicOffsetsStart

A number specifying the offset, in array elements, into dynamicOffsetsData, where the dynamic offset data begins.

dynamicOffsetsLength

A number specifying the number of dynamic offset values to be read from in dynamicOffsetsData.

None (Undefined).

For setBindGroup() calls that use a Uint32Array value for dynamicOffsets, the call will throw with a RangeError DOMException if:

• dynamicOffsetsStart is less than 0.
• dynamicOffsetsStart + dynamicOffsetsLength is greater than dynamicOffsets.length.

The following criteria must be met when calling dispatchWorkgroups(), otherwise a GPUValidationError is generated and the GPUComputePassEncoder becomes invalid:

• index is less than or equal to the GPUDevice's maxBindGroups limit.
• dynamicOffsets.length is the same as the number of entries in bindGroup with hasDynamicOffset: true set.
• For bindGroup entries where the bound buffer's type is "uniform" (see GPUDevice.createBindGroupLayout()), each number in dynamicOffsets is a multiple of the GPUDevice's minUniformBufferOffsetAlignment limit.
• For bindGroup entries where the bound buffer's type is "storage" or "read-only-storage" (see GPUDevice.createBindGroupLayout()), each number in dynamicOffsets is a multiple of the GPUDevice's minStorageBufferOffsetAlignment limit.
• For each bindGroup entry, the bound buffer's offset, plus the corresponding layout entry's minBindingSize, plus the corresponding dynamic offset specified in dynamicOffsets, is less than or equal to the bound buffer's size.

In our basic compute demo, several commands are recorded via a GPUCommandEncoder. Most of these commands originate from the GPUComputePassEncoder created via beginComputePass(). The setBindGroup() call used here is the simplest form, just specifying index and bind group.

const BUFFER_SIZE = 1000;

// ...

// Create GPUCommandEncoder to encode commands to issue to the GPU
const commandEncoder = device.createCommandEncoder();

// Initiate render pass
const passEncoder = commandEncoder.beginComputePass();

// Issue commands
passEncoder.setPipeline(computePipeline);
passEncoder.setBindGroup(0, bindGroup);
passEncoder.dispatchWorkgroups(Math.ceil(BUFFER_SIZE / 64));

// End the render pass
passEncoder.end();

// Copy output buffer to staging buffer
commandEncoder.copyBufferToBuffer(
  output,
  0, // Source offset
  stagingBuffer,
  0, // Destination offset
  BUFFER_SIZE,
);

// End frame by passing array of command buffers to command queue for execution
device.queue.submit([commandEncoder.finish()]);

// ...

• The WebGPU API