Struct SlabAllocator 

pub struct SlabAllocator<Runtime: IsRuntime> { /* private fields */ }

Manages slab allocations and updates over a parameterised buffer.

Create a new instance using SlabAllocator::new.

Upon creation you will need to call SlabAllocator::get_buffer or SlabAllocator::commit at least once before any data is written to the internal buffer.

Implementations

impl SlabAllocator<WgpuRuntime>

pub async fn read( &self, range: impl RangeBounds<usize>, ) -> Result<Vec<u32>, SlabAllocatorError>

Read the slab range from the GPU.

pub async fn read_one<T: SlabItem + Default>( &self, id: Id<T>, ) -> Result<T, SlabAllocatorError>

Read on value from the GPU.

pub async fn read_array<T: SlabItem + Default>( &self, array: Array<T>, ) -> Result<Vec<T>, SlabAllocatorError>

Read an array of typed values from the GPU.

pub fn device(&self) -> &Device

pub fn queue(&self) -> &Queue

impl<R: IsRuntime> SlabAllocator<R>

pub fn new( runtime: impl AsRef<R>, name: impl AsRef<str>, default_buffer_usages: R::BufferUsages, ) -> Self

pub fn len(&self) -> usize

The length of the underlying buffer, in u32 slots.

This does not include data that has not yet been committed.

pub fn is_empty(&self) -> bool

Whether the underlying buffer is empty.

This does not include data that has not yet been committed.

pub fn get_buffer(&self) -> Option<SlabBuffer<R::Buffer>>

Return the internal buffer used by this slab, if it has been created.

pub fn new_value<T: SlabItem + Clone + Send + Sync + 'static>( &self, value: T, ) -> Hybrid<T>

Stage a new value that lives on the GPU and CPU.

pub fn new_array<T: SlabItem + Clone + Send + Sync + 'static>( &self, values: impl IntoIterator<Item = T>, ) -> HybridArray<T>

Stage a contiguous array of new values that live on the GPU and CPU.

pub fn get_updated_source_ids(&self) -> FxHashSet<SourceId>

Return the ids of all sources that require updating.

pub fn has_queued_updates(&self) -> bool

Returns whether any update sources, most likely from Hybrid or Gpu values, have queued updates waiting to be committed.

pub fn commit(&self) -> SlabBuffer<R::Buffer>

Perform upkeep on the slab, synchronizing changes to the internal buffer.

Changes made to Hybrid and Gpu values created by this slab are not committed until this function has been called.

The internal buffer is not created until the first time this function is called.

Returns a SlabBuffer wrapping the internal buffer that is currently in use by the allocator.

pub fn upkeep(&self) -> SlabBuffer<R::Buffer>

👎Deprecated since 0.1.5: please use commit instead

pub fn defrag(&self)

Defragments the internal “recycle” buffer.

pub fn runtime(&self) -> &R

Trait Implementations

impl<R: IsRuntime> Clone for SlabAllocator<R>

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.