Struct SearchsortedPlan 

pub struct SearchsortedPlan<T: Element> { /* private fields */ }

searchsorted plan.

searchsorted(sorted_seq[L], values[N], right) -> output[N] (PyTorch torch.searchsorted). right == false returns lower_bound, right == true returns upper_bound.

When to use: per-query binary search into a sorted 1-D array. Useful for histogram-binning / quantile-bucketing. No BW.

Dtypes: {f32, f64, i32, i64} for both sorted_seq and values; output always i32.

Shape limits: sorted_seq [len_sorted]; values, output [num_queries].

Workspace: none.

Precision guarantee: deterministic, bit-stable. Pure binary search.

Implementations

impl<T: Element> SearchsortedPlan<T>

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

Pick a kernel for desc.

pub fn can_implement(&self, args: &SearchsortedArgs<'_, T>) -> 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 for this plan’s kernel.

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

Launch.