Module gpu 

The GPU context and its asynchronous request.

A scene is set up once the adapter and device exist, and the setup needs a mutable world — so the request is a future the world itself runs, driven by whoever drives the world, and the ready callback lands as a command between two event callbacks. Nothing outside this module needs to know any of that: request_gpu runs the request and calls back when the context is ready.

Re-exports

use std::cell::RefCell; 🔒

use std::rc::Rc; 🔒

use unlit_ecs::Command; 🔒

use unlit_ecs::Entity; 🔒

use unlit_ecs::EntityAccessError; 🔒

use unlit_ecs::World; 🔒

use crate::executor::LocalTasks; 🔒

use unlit_wgpu::capabilities::DeviceCapabilities; 🔒

use unlit_wgpu::capabilities::DeviceTier; 🔒

use unlit3d::frame::FrameBehaviourOrder; 🔒

use unlit3d::frame::OnFrame; 🔒

use unlit3d::winit::builtin::DisplayHandle; 🔒

use unlit3d::winit::builtin::ExitRequest; 🔒

use unlit3d::winit::builtin::Resumed; 🔒

use unlit3d::winit::builtin::window; 🔒

Structs

Gpu

The GPU context a scene draws with.

RunGpuReady 🔒

Run the GPU request’s ready callback with a mutable world.

Enums

GpuRequest 🔒

The answer a GPU request settles on.

GpuRequestError

Why a GPU request could not start the renderer.

RequestStage 🔒

The stages a GPU request moves through, one per frame behaviour.

Constants

REQUEST_GPU_ORDER 🔒

The frame-behaviour order the GPU request runs at: before the default band, so the GPU is there before any behaviour that states no order runs.

Functions

request_adapter 🔒

Ask instance for a high-performance adapter that can present to surface, when one is given.

request_gpu

Start the app’s GPU request once the window exists, and run ready with the context once it lands.