Struct WorkBytes 

pub struct WorkBytes<const N: usize>;

Run in parallel only for an operation that moves at least N bytes.

Byte-reporting operators such as for_each_unit_task_mut_with decide through ExecutionPolicy::parallelize_work. Entry points that report only an element count are treated as moving one byte per element, a lower bound for any element type, so they lean serial rather than over-schedule.

Trait Implementations

impl<const N: usize> Clone for WorkBytes<N>

fn clone(&self) -> WorkBytes<N>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<const N: usize> Copy for WorkBytes<N>

impl<const N: usize> Debug for WorkBytes<N>

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

Formats the value using the given formatter.

impl<const N: usize> Default for WorkBytes<N>

fn default() -> WorkBytes<N>

Returns the “default value” for a type.

impl<const N: usize> ExecutionPolicy for WorkBytes<N>

fn parallelize(len: usize) -> bool

Return true if an operation over len elements should run in parallel.

fn parallelize_work(_len: usize, _chunks: usize, bytes: usize) -> bool

Return true if an operation over len elements, scheduled as chunks tasks and moving bytes in total, should run in parallel.

fn parallelize_chunks(len: usize, chunks: usize) -> bool

Return true if an operation over len elements, partitioned into chunks logical chunks, should run in parallel.

fn parallelize_pair() -> bool

Return true if a fixed two-branch operation should run in parallel.