Struct Gpu 

pub struct Gpu {
    pub instance: Instance,
    pub adapter: Adapter,
    pub device: Device,
    pub queue: Queue,
    pub capabilities: DeviceCapabilities,
}

The GPU context a scene draws with.

Everything here is a wgpu handle, so the handles may cross into another thread’s setup; the scene itself cannot cross threads and is set up where the request started. The native backends’ handles are send-and-sync; the WebGPU backend’s device holds an Rc, so on the web the handles only ever cross tasks, never threads.

Fields

instance: Instance

The instance presentation surfaces are created from.

adapter: Adapter

The adapter the device was requested from.

device: Device

The device every buffer and pipeline belongs to.

queue: Queue

The queue frames are submitted to.

capabilities: DeviceCapabilities

What the device can do beyond the WebGPU baseline, captured while the adapter is still here to report it.

Implementations

impl Gpu

pub async fn request( instance: Instance, surface: Surface<'static>, ) -> Result<Self, Box<dyn Error + Send + Sync>>

Request the adapter and device that can present to surface.

The surface is a probe: it exists only so the adapter is required to present to the window, and is dropped before the device is requested.

pub async fn headless() -> Result<Self, Box<dyn Error + Send + Sync>>

Request an adapter and device without a window to present to, for headless renders.

async fn from_adapter( instance: Instance, adapter: Adapter, ) -> Result<Self, Box<dyn Error + Send + Sync>>

Request a device from adapter and gather the context around it.

Trait Implementations

impl Clone for Gpu

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.