Module burn_wgpu::compute

Compute related module.

Structs

• DynamicKernel
  Wraps a dynamic kernel source into a kernel with launch information such as workgroup.
• StaticKernel
  Wraps a static kernel source into a kernel with launch information such as workgroup.
• WgpuResource
  The memory resource that can be allocated for wgpu.
• WgpuServer
  Wgpu compute server.
• WgpuStorage
  Buffer storage for wgpu.
• WorkGroup
  Provides launch information specifying the number of work groups to be used by a compute shader.

Enums

• WgpuAutotuneKey
  Key for all autotune-enabled operations
• WgpuResourceKind
  How the resource is used, either as a slice or fully.

Traits

• Kernel
  Kernel trait with the source that will be compiled and cached based on the provided id.

Functions

• compute_client
  Get the compute client for the given device.
• init_async
  Init the client async, necessary for wasm.
• select_device
  Select the wgpu device and queue based on the provided device.

Type Aliases

• Server
  Wgpu compute server
• WgpuComputeClient
  Wgpu compute client to communicate with the compute server.
• WgpuHandle
  Wgpu server handle.