Struct CpuExecution 

pub struct CpuExecution { /* private fields */ }

CPU execution context with runtime ISA detection and, with parallel, a local pool of exactly Self::thread_count workers.

This type deliberately implements no geometry operation trait. A provider only implements MeshBoolean, MeshCompiler, or another capability after the algorithm exists and is verified.

Implementations

impl CpuExecution

pub fn portable() -> CpuExecution

Always-available portable execution context.

pub fn detect() -> CpuExecution

Best instruction path compiled into this binary and available at runtime.

pub const fn instruction_set(&self) -> CpuInstructionSet

Runtime-selected instruction set.

pub const fn features(&self) -> CpuFeatures

Detected machine features.

pub const fn thread_count(&self) -> NonZero<usize>

Configured worker bound.

pub fn install<R>(&self, operation: impl FnOnce() -> R + Send) -> R

where R: Send,

Execute one closure inside this context’s local Rayon pool, which has exactly Self::thread_count workers, one included: inside it rayon::current_num_threads() is that count, whatever pool the caller runs in.

Panics

For a context from Self::portable or Self::detect, the pool is started on first use, and this panics if the operating system refuses the thread, as a first use of Rayon’s global pool does. A context from crate::CpuExecutionBuilder started its pool when it was built and reported a failure there.

Trait Implementations

impl Clone for CpuExecution

fn clone(&self) -> CpuExecution

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CpuExecution

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

Formats the value using the given formatter.