Struct GpuPirInterner 

pub struct GpuPirInterner { /* private fields */ }

GPU PIR interner (device-side).

Implementations

impl GpuPirInterner

pub fn new( provider: &Arc<CudaKernelProvider>, node_cap: u32, child_cap: u32, ) -> Result<Self>

Allocates an interner with fixed node and child capacities.

pub fn graph(&self) -> &GpuPirGraph

Returns the device-resident interned graph.

pub fn intern_batch( &mut self, batch: &PirBatch, ) -> Result<TrackedCudaSlice<u32>>

Uploads and interns a host batch, returning canonical node identifiers.

pub fn intern_device_batch( &mut self, batch: &mut GpuPirBatch, ) -> Result<TrackedCudaSlice<u32>>

Interns a device-resident batch and returns canonical node identifiers.