Skip to main content

TriuPlan

Struct TriuPlan 

Source
pub struct TriuPlan<T: Element, const N: usize> { /* private fields */ }
Expand description

triu plan.

y = torch.triu(x, diagonal) — upper-triangular mask on the last two dims of x.

When to use: forward triu. Pair with TriuBackwardPlan — the BW is the same mask applied to d_output.

Dtypes: {f16, bf16, f32, f64, i32, i64, Bool}.

Shape limits: rank in [2, 8]; last two dims (M = shape[N-2], N_cols = shape[N-1]) define the matrix; everything before is the batch prefix.

Workspace: none.

Precision guarantee: deterministic, bit-stable, bit-exact — pure element select + zero, no arithmetic.

Implementations§

Source§

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

Source

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

Pick a kernel for desc.

Source

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

Validate args.

Source

pub fn workspace_size(&self) -> usize

Workspace size in bytes. Always 0.

Source

pub fn sku(&self) -> KernelSku

Identity of the kernel this plan picked.

Source

pub fn precision_guarantee(&self) -> PrecisionGuarantee

Numerical guarantees.

Source

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

Launch.

Dispatch policy: if both input and output are canonical row-major contiguous, route to the contig fast path (baracuda_kernels_triu_<dtype>_run). Otherwise route to the strided sibling (baracuda_kernels_triu_<dtype>_strided_run, Phase 14.3) which threads per-axis signed strides for input and output through the kernel parameter block.

Auto Trait Implementations§

§

impl<T, const N: usize> Freeze for TriuPlan<T, N>

§

impl<T, const N: usize> RefUnwindSafe for TriuPlan<T, N>
where T: RefUnwindSafe,

§

impl<T, const N: usize> Send for TriuPlan<T, N>
where T: Send,

§

impl<T, const N: usize> Sync for TriuPlan<T, N>
where T: Sync,

§

impl<T, const N: usize> Unpin for TriuPlan<T, N>
where T: Unpin,

§

impl<T, const N: usize> UnsafeUnpin for TriuPlan<T, N>

§

impl<T, const N: usize> UnwindSafe for TriuPlan<T, N>
where T: UnwindSafe,

Blanket Implementations§

Source§

impl<T> Any for T
where T: 'static + ?Sized,

Source§

fn type_id(&self) -> TypeId

Gets the TypeId of self. Read more
Source§

impl<T> Borrow<T> for T
where T: ?Sized,

Source§

fn borrow(&self) -> &T

Immutably borrows from an owned value. Read more
Source§

impl<T> BorrowMut<T> for T
where T: ?Sized,

Source§

fn borrow_mut(&mut self) -> &mut T

Mutably borrows from an owned value. Read more
Source§

impl<T> From<T> for T

Source§

fn from(t: T) -> T

Returns the argument unchanged.

Source§

impl<T, U> Into<U> for T
where U: From<T>,

Source§

fn into(self) -> U

Calls U::from(self).

That is, this conversion is whatever the implementation of From<T> for U chooses to do.

Source§

impl<T, U> TryFrom<U> for T
where U: Into<T>,

Source§

type Error = Infallible

The type returned in the event of a conversion error.
Source§

fn try_from(value: U) -> Result<T, <T as TryFrom<U>>::Error>

Performs the conversion.
Source§

impl<T, U> TryInto<U> for T
where U: TryFrom<T>,

Source§

type Error = <U as TryFrom<T>>::Error

The type returned in the event of a conversion error.
Source§

fn try_into(self) -> Result<U, <U as TryFrom<T>>::Error>

Performs the conversion.