Enum ConstValue 

pub enum ConstValue {
    Int(i64),
    UInt(u64),
    Float(f64),
    Bool(bool),
}

Constant value that can be stored in a UOp.

Variants

Int(i64)

UInt(u64)

Float(f64)

Bool(bool)

Implementations

impl ConstValue

pub const fn dtype(&self) -> DType

pub const fn zero(dtype: ScalarDType) -> Self

pub const fn one(dtype: ScalarDType) -> Self

pub const fn neg_one(dtype: ScalarDType) -> Option<Self>

pub const fn min(dtype: ScalarDType) -> Self

Minimum representable value for a scalar dtype (matches Tinygrad’s dtypes.min).

pub const fn max(dtype: ScalarDType) -> Self

Maximum representable value for a scalar dtype (matches Tinygrad’s dtypes.max).

pub fn cast(&self, dtype: &DType) -> Option<Self>

Cast this constant value to the target dtype.

Returns None if:

• The target dtype is not a scalar type
• The target dtype is not representable as a ConstValue (e.g., Void, Index, special float formats)

Safety and Semantics

This method performs constant folding for cast operations and allows ALL casts (including lossy ones like float->int) since the user explicitly wrote the cast operation.

Uses Rust’s as operator for conversions, which follows C semantics:

• Truncation for narrowing conversions (e.g., i64 -> i32)
• Wrap-around for unsigned overflow
• Truncation toward zero for float-to-int conversions

For multi-stage conversions (e.g., casting through intermediate types), the value is cast to the target width and then extended back to the storage type. Example: i64 -> i8 -> i64 ensures proper sign extension.

pub const fn is_zero(&self) -> bool

Returns true if this constant is zero (additive identity).

Works for all numeric types: Int, UInt, Float, Bool.

pub const fn is_one(&self) -> bool

Returns true if this constant is one (multiplicative identity).

Works for all numeric types: Int, UInt, Float, Bool.

pub const fn is_neg_one(&self) -> bool

Returns true if this constant is negative one.

Used for patterns like x // -1 → -x.

pub const fn try_int(&self) -> Option<i64>

Try to extract an integer value (i64 or u64 as i64).

Used for constant pattern matching with specific integer values.

pub const fn try_float(&self) -> Option<f64>

Try to extract a float value (f64).

Used for constant pattern matching with specific float values.

pub fn truncate(self, dtype: ScalarDType) -> Self

Truncate value to fit within dtype boundaries (two’s complement wrapping).

This is equivalent to Tinygrad’s ctypes-based truncation. Used for constant folding to ensure results respect the target dtype’s bit width.

Trait Implementations

impl Clone for ConstValue

fn clone(&self) -> ConstValue

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ConstValue

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ConstValue

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl From<bool> for ConstValue

fn from(value: bool) -> Self

Converts to this type from the input type.

impl From<f32> for ConstValue

fn from(v: f32) -> Self

Converts to this type from the input type.

impl From<f64> for ConstValue

fn from(value: f64) -> Self

Converts to this type from the input type.

impl From<i16> for ConstValue

fn from(v: i16) -> Self

Converts to this type from the input type.

impl From<i32> for ConstValue

fn from(v: i32) -> Self

Converts to this type from the input type.

impl From<i64> for ConstValue

fn from(value: i64) -> Self

Converts to this type from the input type.

impl From<i8> for ConstValue

fn from(v: i8) -> Self

Converts to this type from the input type.

impl From<u16> for ConstValue

fn from(v: u16) -> Self

Converts to this type from the input type.

impl From<u32> for ConstValue

fn from(v: u32) -> Self

Converts to this type from the input type.

impl From<u64> for ConstValue

fn from(value: u64) -> Self

Converts to this type from the input type.

impl From<u8> for ConstValue

fn from(v: u8) -> Self

Converts to this type from the input type.

impl Hash for ConstValue

Manual Hash impl because f64 doesn’t implement Hash. Uses to_bits() for floats, which means NaN values with identical bit patterns hash equally.

fn hash<H: Hasher>(&self, state: &mut H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl IntoUOp for ConstValue

fn into_uop(self, dtype: DType) -> Arc<UOp>

impl PartialEq for ConstValue

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

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Copy for ConstValue

impl StructuralPartialEq for ConstValue