Enum ReduceFunct 

#[repr(u8)]
pub enum ReduceFunct {
    Fadd = 0,
    Fmax = 2,
    Fmin = 3,
    Add = 4,
    Max = 6,
    Min = 7,
    Move = 8,
}

Reduction function selector for tensor_recv.

Specifies how the received values are combined with the values already held in the destination FP registers. (PRM Table 9-8, FUNCT field.)

Variants

Fadd = 0

C[i] = C[i] + src[i] (fp32 addition)

Fmax = 2

C[i] = fmax(C[i], src[i])

Fmin = 3

C[i] = fmin(C[i], src[i])

Add = 4

C[i] = C[i] + src[i] (integer addition on bit pattern)

Max = 6

C[i] = max(C[i], src[i]) (signed 32-bit integer comparison)

Min = 7

C[i] = min(C[i], src[i]) (signed 32-bit integer comparison)

Move = 8

C[i] = src[i] (unconditional move)

Trait Implementations

impl Clone for ReduceFunct

fn clone(&self) -> ReduceFunct

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for ReduceFunct

impl Debug for ReduceFunct

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

Formats the value using the given formatter.

impl Eq for ReduceFunct

impl PartialEq for ReduceFunct

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for ReduceFunct