GPUDevice: createComputePipelineAsync() method

Limited availability

This feature is not Baseline because it does not work in some of the most widely-used browsers.

Experimental: This is an experimental technology
Check the Browser compatibility table carefully before using this in production.

Secure context: This feature is available only in secure contexts (HTTPS), in some or all supporting browsers.

Note: This feature is available in Web Workers.

The createComputePipelineAsync() method of the GPUDevice interface returns a Promise that fulfills with a GPUComputePipeline, which can control the compute shader stage and be used in a GPUComputePassEncoder, once the pipeline can be used without any stalling.

Note: It is generally preferable to use this method over GPUDevice.createComputePipeline() whenever possible, as it prevents blocking of GPU operation execution on pipeline compilation.

Syntax

createComputePipeline(descriptor)

Parameters

descriptor: See the descriptor definition for the GPUDevice.createComputePipeline() method.

Return value

A Promise that fulfills with a GPUComputePipeline object instance when the created pipeline is ready to be used without additional delay.

Validation

If pipeline creation fails and the resulting pipeline becomes invalid as a result, the returned promise rejects with a GPUPipelineError:

If this is due to an internal error, the GPUPipelineError will have a reason of "internal".
If this is due to a validation error, the GPUPipelineError will have a reason of "validation".

A validation error can occur if any of the following are false:

The workgroup storage size used by the module referenced inside the compute property is less than or equal to the GPUDevice's maxComputeWorkgroupStorageSize limit.
The module uses a number of compute invocations per workgroup less than or equal to the GPUDevice's maxComputeInvocationsPerWorkgroup limit.
The module's workgroup size is less than or equal to the GPUDevice's corresponding maxComputeWorkgroupSizeX, maxComputeWorkgroupSizeY, or maxComputeWorkgroupSizeZ limit.
If the entryPoint property is omitted, the shader code contains a single compute shader entry point function for the browser to use as the default entry point.

Examples

Note: The WebGPU samples feature many more examples.

Basic example

The following example shows a process of:

Creating a bind group layout with GPUDevice.createBindGroupLayout().
Feeding the bindGroupLayout into GPUDevice.createPipelineLayout() to create a GPUPipelineLayout.
Using that value immediately in a createComputePipelineAsync() call to create a GPUComputePipeline.

async function init() {
  // ...

  const bindGroupLayout = device.createBindGroupLayout({
    entries: [
      {
        binding: 0,
        visibility: GPUShaderStage.COMPUTE,
        buffer: {
          type: "storage",
        },
      },
    ],
  });

  const computePipeline = await device.createComputePipelineAsync({
    layout: device.createPipelineLayout({
      bindGroupLayouts: [bindGroupLayout],
    }),
    compute: {
      module: shaderModule,
      entryPoint: "main",
    },
  });

  // ...
}

Specifications

Specification
WebGPU # dom-gpudevice-createcomputepipelineasync

Browser compatibility

BCD tables only load in the browser