Enum GpuMixedPrecisionPolicy 

pub enum GpuMixedPrecisionPolicy {
    Off,
    Refinement,
    Never,
}

Variants

Off

Always use fp64 factorization; no refinement attempted.

Refinement

Attempt fp32 Cholesky factorization followed by up to REFINEMENT_MAX_STEPS fp64-residual refinement steps. Policy admits the attempt only when p ≥ REFINEMENT_MIN_P (so that the fp64 GEMV overhead is amortized) and the measured residual drops monotonically. Falls back to fp64 factorization automatically when the residual does not decrease (κ(A)·u ≥ 1 regime) or when the fp32 POTRF itself fails.

Never

Always use fp64 factorization; equivalent to Off but signals that an explicit policy decision was taken.

Trait Implementations

impl Clone for GpuMixedPrecisionPolicy

fn clone(&self) -> GpuMixedPrecisionPolicy

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for GpuMixedPrecisionPolicy

impl Debug for GpuMixedPrecisionPolicy

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for GpuMixedPrecisionPolicy

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for GpuMixedPrecisionPolicy

impl PartialEq for GpuMixedPrecisionPolicy

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for GpuMixedPrecisionPolicy

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for GpuMixedPrecisionPolicy