Enum MixedPrecision 

pub enum MixedPrecision {
    F16,
    Bf16,
}

Reduced-precision input storage format for a mixed-precision GEMM.

Mixed-precision GEMM stores its A/B operands in a 16-bit format but accumulates the dot products in f32 (the Tensor-Core / WMMA contract). This enum names the input storage format; the accumulator and output are always f32.

Variants

F16

IEEE-754 binary16 (half): 1 sign, 5 exponent, 10 mantissa bits.

Bf16

Google bfloat16: 1 sign, 8 exponent, 7 mantissa bits (the top 16 bits of an f32).

Trait Implementations

impl Clone for MixedPrecision

fn clone(&self) -> MixedPrecision

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for MixedPrecision

impl Debug for MixedPrecision

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

Formats the value using the given formatter.

impl Display for MixedPrecision

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

Formats the value using the given formatter.

impl Eq for MixedPrecision

impl Hash for MixedPrecision

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for MixedPrecision

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for MixedPrecision