Struct Gpu 

pub struct Gpu {
    pub device: Device,
    pub queue: Queue,
}

Handle to a GPU device

Fields

device: Device

GPU device

queue: Queue

GPU queue

Implementations

impl Gpu

pub async fn init() -> Result<Gpu, InitError>

Returns a Gpu object with customized settings

Non-default settings are as follows:

• We request a wgpu::PowerPreference::HighPerformance adapter
• We enable the wgpu::Features::TIMESTAMP_QUERY feature

This is a helper function for simplicity; more sophisticated systems will likely construct the adapter, device, and queue themselves.

pub fn read_buffer_for<T>(&self, buf: &FlexBuffer<T>) -> ReadBuffer<T>

where T: BufferTag,

Returns a readable buffer for the given buffer

See copy and map for how to use the resulting read buffer.

pub fn read_buffer<T>(&self, name: &str) -> ReadBuffer<T>

where T: BufferTag, T::S: TryFrom<u32>,

Builds a new readable buffer

The buffer has an arbitrary starting size and will be resized when used in copy.

If the target buffer is known, use read_buffer_for to avoid this reallocation.

pub fn read_vec<T: BufferTag>(&self, buf: &FlexBuffer<T>) -> Vec<T::T>

where T::T: FromBytes + Immutable + Clone + Copy,

Helper function to read from a buffer to a Vec

Under the hood, this function simply calls read_buffer_for, copy, map, and to_vec.

Note that this function allocates a GPU buffer; if the same source buffer is being read repeatedly, it’s recommended to build the read buffer once and then call copy, map, to_vec repeatedly.

pub fn copy<A>(&self, src: &FlexBuffer<A>, dst: &mut ReadBuffer<A>)

where A: BufferTag,

Copies from a GPU-resident buffer to a host-mappable buffer

The host-mappable destination buffer is resized to fit the data.

pub fn encode_copy<A>( &self, src: &FlexBuffer<A>, dst: &mut ReadBuffer<A>, encoder: &mut CommandEncoder, )

where A: BufferTag,

Low-level command to submit a buffer copy to a command encoder

See copy for details

pub fn map<'a, T>(&self, image: &'a mut ReadBuffer<T>) -> MappedBuffer<'a, T>

where T: BufferTag, T::T: Immutable + FromBytes + Copy,

Blocking function to build a new mapped buffer

pub async fn map_async<'a, T>( &self, data: &'a mut ReadBuffer<T>, ) -> MappedBuffer<'a, T>

where T: BufferTag, T::T: Immutable + FromBytes + Copy,

Async function to build a new mapped buffer

This can be called on either native or web platforms. On native platforms, the single await is trivial (guaranteed to always be ready); on the web, the mapping sends us back to the event loop until it’s ready.

pub fn map_image<'a, T>( &self, data: &'a mut ReadBuffer<T>, ) -> MappedImage<'a, T>

where T: BufferTag, T::T: Immutable + FromBytes + Copy, T::S: RenderSize,

Blocking function to build a new mapped image

pub async fn map_image_async<'a, T>( &self, data: &'a mut ReadBuffer<T>, ) -> MappedImage<'a, T>

where T: BufferTag, T::T: Immutable + FromBytes + Copy, T::S: RenderSize,

Async function to build a new mapped image

This can be called on either native or web platforms. On native platforms, the single await is trivial (guaranteed to always be ready); on the web, the mapping sends us back to the event loop until it’s ready.

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.