Floating_Point_Type

class Floating_Point_Type (abstract, logical)

This is the superclass for all C floating point types.

Direct parents:

• inherits Basic_Type

Ancestor classes and implemented interfaces:

Basic_Type, Declarative_Node, Logical_IR_Root, Member, Named_Entity, Named_Type, Symbol, Type

Direct child classes:

• subclass C_Long_Double_Type

• subclass C_Double_Type

• subclass C_Float_Type

• subclass C_Float_Fp16_Type

• subclass C_Float_Bf16_Type

• subclass Interchange_Float16_Type

• subclass Interchange_Float32_Type

• subclass Interchange_Float64_Type

• subclass Interchange_Float128_Type

• subclass Extended_Float32_Type

• subclass Extended_Float64_Type

• subclass C_Float80_Type

• subclass C_Float128_Type

Field	Kind	Type	Description
Mantissa_Digits	attribute	builtin Integer	Precision of the mantissa, in bits. Includes the implicit mantissa bit which is not actually stored. In the range 0.5 to 1.0, the step size between neighboring floats is 2**(-Mantissa_Digits). Corresponds to std::numeric_limits<T>::digits.
Min_Exponent	attribute	builtin Integer	One more than the smallest negative power of 2 that is a valid normalized floating-point value. Corresponds to std::numeric_limits<T>::min_exponent.
Max_Exponent	attribute	builtin Integer	One more than the largest integer power of 2 that is a valid finite floating-point value. Corresponds to std::numeric_limits<T>::max_exponent.
Size (from Type)	attribute	builtin Natural
Alignment_Log (from Type)	attribute	builtin Byte	log2 of the type’s alignment
Alignment (from Type)	attribute	builtin Natural	Alignment of the type (must be a power of 2)
Size_In_Bits (from Type)	attribute	builtin UInt64	Size of the type, in bits.
Parent (from Logical_IR_Root)	semantic	class Logical_IR_Root
Physical (from Logical_IR_Root)	semantic to physical	class Physical_IR_Root
Visibility (from Member)	attribute	enum Visibility_Type
Name (from Named_Entity)	attribute	builtin String