Module compute

A unified system for GPU-based computation.

This module provides a structured way to define and dispatch compute shaders as part of a sequential rendering and computation workflow. It integrates seamlessly with the unified command system to enable mixed graphics and compute workloads.

Key Components

• ComputeCommand: A trait marking a command as a compute operation with optional barrier support.
• ComputablePipeline: A trait for a specific compute task that processes a command within a wgpu::ComputePass.
• ComputePipelineRegistry: The central dispatcher that manages all registered pipelines.

Workflow

1. Define a Command: Create a struct that implements ComputeCommand.
2. Implement a Pipeline: Create a struct that implements ComputablePipeline<YourCommand>. This involves setting up the wgpu::ComputePipeline and defining the dispatch method.
3. Register the Pipeline: During application startup, register an instance of your pipeline with the ComputePipelineRegistry.
4. Submit Commands: Components submit ComputeCommands, which are then dispatched by the renderer to the appropriate pipeline through the unified command system.

Barrier Support

Compute commands can specify barrier requirements to ensure proper synchronization with previous rendering operations, enabling post-processing effects and multi-pass algorithms.

Re-exports

pub use command::ComputeCommand;

pub use pipeline::ComputablePipeline;

pub use pipeline::ComputePipelineRegistry;

pub use resource::ComputeResource;

pub use resource::ComputeResourceManager;

pub use resource::ComputeResourceRef;

Modules

command

Compute command trait and related types.

pipeline

GPU compute pipeline system for Tessera UI framework.

resource