Struct PirlsLoopAdmission 

pub struct PirlsLoopAdmission {
    pub n: usize,
    pub p: usize,
    pub family: Option<PirlsLoopFamilyKind>,
    pub curvature: PirlsLoopCurvatureKind,
    pub gpu_available: bool,
}

Inputs to [should_use_gpu_pirls_loop]. Each field comes from data the CPU PIRLS entry has on hand before it touches the eigendecomposition engine, so the admission check itself is allocation-free and can short- circuit before any heavy work happens.

Fields

n: usize

Number of rows in the active (post-transform) design matrix.

p: usize

Number of columns in the active design (i.e. p of Xᵀ X).

family: Option<PirlsLoopFamilyKind>

Some(_) when the inner family maps onto one of the six JIT-cached PirlsRowFamily variants; None for custom families that still require Stage 6 Level B and have not yet been admitted here.

curvature: PirlsLoopCurvatureKind

Curvature surface the inner loop will use; the GPU loop has Fisher + Observed kernels, anything else (e.g. expected-projection surrogates) is not admitted.

gpu_available: bool

True when the CUDA runtime is initialised on this host (i.e. GpuRuntime::global().is_some()).

Trait Implementations

impl Clone for PirlsLoopAdmission

fn clone(&self) -> PirlsLoopAdmission

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for PirlsLoopAdmission

impl Debug for PirlsLoopAdmission

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

Formats the value using the given formatter.