Enum WgpuKernelOp 

pub enum WgpuKernelOp {
    Add,
    Sub,
    Mul,
    Div,
    Neg,
    Sqrt,
    Exp,
    Log,
    Sin,
    Cos,
    Sum,
    Min,
    Max,
    Norm,
    Transpose,
    Dot,
    Matmul,
}

Portable operation implemented by a wgpu tensor kernel.

Variants

Add

Element-wise addition.

Sub

Element-wise subtraction.

Mul

Element-wise multiplication.

Div

Element-wise division.

Neg

Element-wise negation.

Sqrt

Element-wise square root.

Exp

Element-wise exponential.

Log

Element-wise natural logarithm.

Sin

Element-wise sine.

Cos

Element-wise cosine.

Sum

Whole-tensor sum reduction.

Min

Whole-tensor minimum reduction.

Max

Whole-tensor maximum reduction.

Norm

Whole-tensor Euclidean norm.

Transpose

Matrix transpose.

Dot

Vector dot product.

Matmul

Vector or matrix multiplication.

Implementations

impl WgpuKernelOp

pub fn is_binary(self) -> bool

Returns true when the kernel consumes two input tensors.

pub fn is_unary(self) -> bool

Returns true when the kernel consumes exactly one input tensor.

pub fn is_reduction(self) -> bool

Returns true when the kernel performs fixed-tree accumulation.

pub fn is_linalg(self) -> bool

Returns true when the kernel performs a linalg memory/product primitive.

Trait Implementations

impl Clone for WgpuKernelOp

fn clone(&self) -> WgpuKernelOp

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for WgpuKernelOp

impl Debug for WgpuKernelOp

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

Formats the value using the given formatter.

impl Eq for WgpuKernelOp

impl PartialEq for WgpuKernelOp

fn eq(&self, other: &WgpuKernelOp) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for WgpuKernelOp