Struct RepeatPlan 

pub struct RepeatPlan<T: Element, const N: usize> { /* private fields */ }

repeat plan.

Per-axis tile: output.shape[d] = input.shape[d] * repeats[d] (PyTorch torch.repeat). Kernel walks output cells and computes input coords as output_coord[d] % input.shape[d].

When to use: forward repeat. Pair with RepeatBackwardPlan.

Dtypes: {f32, f64, f16, bf16}. Pure load + store.

Shape limits: rank in [1, 8]; repeats[d] ≥ 1.

Workspace: none.

Precision guarantee: deterministic, bit-stable, bit-exact.

Implementations

impl<T: Element, const N: usize> RepeatPlan<T, N>

pub fn select( _stream: &Stream, desc: &RepeatDescriptor<N>, _pref: PlanPreference, ) -> Result<Self>

Pick a kernel for desc.

pub fn can_implement(&self, args: &RepeatArgs<'_, T, N>) -> Result<()>

Validate args.

pub fn workspace_size(&self) -> usize

Workspace size in bytes.

pub fn sku(&self) -> KernelSku

Identity of the kernel this plan picked.

pub fn precision_guarantee(&self) -> PrecisionGuarantee

Numerical guarantees.

pub fn run( &self, stream: &Stream, _workspace: Workspace<'_>, args: RepeatArgs<'_, T, N>, ) -> Result<()>

Launch.