Struct nannou::wgpu::RowPaddedBuffer

pub struct RowPaddedBuffer { /* fields omitted */ }

A wrapper around a wgpu buffer suitable for copying to and from Textures. Automatically handles the padding necessary for buffer-to-texture and texture-to-buffer copies.

Note: as of wgpu 0.6, texture-to-buffer and buffer-to-texture copies require that image rows are padded to a multiple wgpu::COPY_BYTES_PER_ROW_ALIGNMENT bytes. Note that this is a requirement on the buffers, not on the textures! You can have textures of whatever size you like, but when you copy them to/from a buffer, the buffer rows need padding. This is referred to as “pitch alignment”.

In a RowPaddedBuffer, the image is stored in row-major order, with rows padded at the end with uninitialized bytes to reach the necessary size.

Implementations

impl RowPaddedBuffer

pub fn new(
    device: &Device,
    width: u32,
    height: u32,
    usage: BufferUsage
) -> RowPaddedBuffer

Create a row-padded buffer on the device.

Width should be given in bytes.

pub fn from_image_buffer<P, Container>(
    device: &Device,
    image_buffer: &ImageBuffer<P, Container>
) -> RowPaddedBuffer where
    P: 'static + Pixel,
    Container: Deref<Target = [P::Subpixel]>, 

Initialize from an image buffer (i.e. an image on CPU).

pub fn for_texture(
    device: &Device,
    texture: &Texture,
    usage: BufferUsage
) -> RowPaddedBuffer

Creates a buffer compatible with a 2d slice of the given texture.

pub fn width(&self) -> u32

The width of the buffer, in bytes, NOT including padding bytes.

pub fn row_padding(&self) -> u32

The padding of each row of this buffer.

pub fn padded_width(&self) -> u32

The width of the buffer, in bytes, INCLUDING padding bytes.

pub fn height(&self) -> u32

The height of the buffer.

pub fn write(&self, buf: &[u8])

Copy data into the padded buffer.

Will copy data_width bytes of data into each row of the buffer, leaving the remainder of the buffer unmodified.

The buffer usage must include BufferUsage::map_read().

pub async fn read<'b>(
    &'b self
) -> Result<ImageReadMapping<'b>, BufferAsyncError>

Asynchronously maps the buffer of bytes from GPU to host memory and, once mapped, calls the given user callback with the data represented as an ImageReadMapping.

Note: The given callback will not be called until the memory is mapped and the device is polled. You should not rely on the callback being called immediately.

pub fn encode_copy_into(
    &self,
    encoder: &mut CommandEncoder,
    destination: &Texture
)

Encode a copy into a texture. Assumes the texture is 2d.

The copy will not be performed until the encoded command buffer is submitted.

pub fn encode_copy_into_at(
    &self,
    encoder: &mut CommandEncoder,
    destination: &Texture,
    depth: u32
)

Encode a copy into a 3d texture at a given depth. Will copy this buffer (modulo padding) to a slice of the texture at the given depth.

The copy will not be performed until the encoded command buffer is submitted.

pub fn encode_copy_from(&self, encoder: &mut CommandEncoder, source: &Texture)

Encode a copy from a texture.

The copy will not be performed until the encoded command buffer is submitted.

pub fn encode_copy_from_at(
    &self,
    encoder: &mut CommandEncoder,
    source: &Texture,
    depth: u32
)

Encode a copy from a 3d texture at a given depth. Will copy a slice of the texture to fill this whole buffer (modulo padding).

The copy will not be performed until the encoded command buffer is submitted.

Trait Implementations

impl Debug for RowPaddedBuffer

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.