Struct CalibrationPoint 

pub struct CalibrationPoint {

    pub id: String,
    pub shape: (usize, usize, usize),
    pub chip: String,
    pub stack: String,
    pub program: String,
    pub strategy: String,
    pub knobs: String,
    pub fidelity: Fidelity,
    pub fp32: bool,
    pub l1acc: bool,
    pub exact: bool,
    pub plan: PlanDesc,
    pub cycles: u64,
    pub work: Option<MeasuredWork>,
    pub pcc: f64,
    pub pcc_floor: f64,
    pub reproduce: String,
    pub provenance: String,
}

One ttsim ledger row as a calibration point.

Fields

id: String

shape: (usize, usize, usize)

chip: String

stack: String

program: String

strategy: String

knobs: String

fidelity: Fidelity

fp32: bool

l1acc: bool

exact: bool

fp32 packing ⇔ the program kept its partials exact.

plan: PlanDesc

cycles: u64

work: Option<MeasuredWork>

What the machine actually did, when the row was measured on the instrumented simulator.

★ THIS, NOT cycles, IS WHAT RANKS TWO MEASURED PLANS. cycles is ttsim’s bracket counter and ttsim_instrument_audit.json shows it is not a function of the program — eight times the reduction depth moves it 0.08%. A row carries it for provenance; nothing may order plans by it.

pcc: f64

pcc_floor: f64

reproduce: String

provenance: String

Trait Implementations

impl Clone for CalibrationPoint

fn clone(&self) -> CalibrationPoint

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CalibrationPoint

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for CalibrationPoint

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for CalibrationPoint

fn eq(&self, other: &CalibrationPoint) -> 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 CalibrationPoint

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 CalibrationPoint