Struct MappedTensorReader

pub struct MappedTensorReader<EG: Numeric> {
    pub tensor: TensorMap<EG>,
    pub tile_x: u32,
    pub tile_y: u32,
    pub batch: u32,
}

A view of a tensor that starts reading data from a specified offset. Uses a TensorMap to actually execute the load.

Fields

tensor: TensorMap<EG>

tile_x: u32

tile_y: u32

batch: u32

Implementations

impl<EG: Numeric> MappedTensorReader<EG>

pub fn new(tensor: TensorMap<EG>, tile_x: u32, tile_y: u32, batch: u32) -> Self

Instantiate a read view over the given tensor, pre-fetching needed strides and shapes

pub fn update_view(&mut self, k_offset: u32, ident: Ident)

Advance the view along the k dimension by a specified offset, k_offset.

pub fn __expand_new( context: &mut Scope, tensor: <TensorMap<EG> as CubeType>::ExpandType, tile_x: <u32 as CubeType>::ExpandType, tile_y: <u32 as CubeType>::ExpandType, batch: <u32 as CubeType>::ExpandType, ) -> <Self as CubeType>::ExpandType

pub fn __expand_update_view( context: &mut Scope, this: <Self as CubeType>::ExpandType, k_offset: <u32 as CubeType>::ExpandType, ident: Ident, ) -> <() as CubeType>::ExpandType

Trait Implementations

impl<EG: Numeric> CubeType for MappedTensorReader<EG>

type ExpandType = MappedTensorReaderExpand<EG>

fn init(scope: &mut Scope, expand: Self::ExpandType) -> Self::ExpandType

Wrapper around the init method, necessary to type inference.

impl<EG: Numeric> Send for MappedTensorReader<EG>

impl<EG: Numeric> Sync for MappedTensorReader<EG>