Struct BindGroupCache 

pub struct BindGroupCache { /* private fields */ }

Bounded LRU cache for wgpu bind groups, keyed by layout identity and the ordered set of buffer handles bound to that layout.

wgpu bind-group creation is non-trivial; this cache eliminates the redundant cost on repeated dispatches that share the same buffer handles. Capped at 256 entries with LRU eviction to prevent descriptor-heap exhaustion on long-running servers.

Implementations

impl BindGroupCache

pub fn new() -> Self

Create a bind-group cache with the default 256-entry cap.

pub fn with_cap(cap: usize) -> Self

Create with an explicit cap (used by tests and consumers that want to size the LRU against known working-set bounds).

pub fn get_or_create( &self, layout_id: usize, handles: &[GpuBufferHandle], factory: impl FnOnce() -> BindGroup, ) -> Arc<BindGroup>

Return a cached bind group or create one with factory.

layout_id must uniquely identify the wgpu::BindGroupLayout (e.g. Arc::as_ptr(layout).addr()). handles must be in the same order as the wgpu::BindGroupEntry slice that the caller will pass to create_bind_group so that identical handle sets map to the same cache key.

pub fn stats(&self) -> BindGroupCacheStats

Return cache statistics for diagnostics and tests.

Trait Implementations

impl Clone for BindGroupCache

fn clone(&self) -> BindGroupCache

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for BindGroupCache

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

Formats the value using the given formatter.

impl Default for BindGroupCache

fn default() -> Self

Returns the “default value” for a type.