Enum WorkerPlacement 

pub enum WorkerPlacement {
    Unbound,
    Pinned,
}

Whether scheduler workers are confined to logical processors.

The policy is an opt-in startup contract. It changes what the runtime enforces, and therefore what it may publish about worker placement.

Variants

Unbound

Workers run wherever the operating system schedules them.

The runtime claims no worker-to-processor or worker-to-node assignment, and the topology-aware victim tier stays inactive.

Pinned

Worker i is confined to one logical processor for its whole life.

Processors are taken from the detected topology in ascending id order, worker i receiving processor i % processor_count, so a pool larger than the machine shares processors. Construction fails closed: when the topology cannot be detected, or the operating system refuses any worker’s binding, the scheduler is not returned and the error names the lowest-numbered worker that failed. On a host with more than one NUMA node the bound assignment is what the runtime publishes for topology-aware stealing.

Trait Implementations

impl Clone for WorkerPlacement

fn clone(&self) -> WorkerPlacement

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for WorkerPlacement

impl Debug for WorkerPlacement

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

Formats the value using the given formatter.

impl Default for WorkerPlacement

fn default() -> WorkerPlacement

Returns the “default value” for a type.

impl Eq for WorkerPlacement

impl PartialEq for WorkerPlacement

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for WorkerPlacement