Ocean::Float16 Class Reference

Definition of a 16-bit float type. More...

Public Member Functions
	Float16 ()=default
	Default constructor. More...
	Float16 (const uint16_t sign, const uint16_t fraction, const uint16_t exponent)
	Creates a new 16-bit float. More...
	Float16 (const uint16_t binary)
	Creates a new 16-bit float based on a 16-bit binary value. More...
uint16_t	binary () const
	Returns the binary representation of this 16-bit float value. More...
	operator float () const
	Cast operator to 32-bit float. More...
Float16	operator- () const
	Negates the 16-bit float value. More...
bool	operator== (const Float16 &second) const
	Returns whether two 16-bit float values are identical. More...
bool	operator!= (const Float16 &second) const
	Returns whether two 16-bit float values are not identical. More...

Static Public Member Functions
static Float16	infinity ()
	Returns +infinity. More...

Protected Attributes
union {
uint16_t   binary_ = 0u
	The binary representation. More...
struct {
uint16_t   fraction_: 10
	The mantissa composed of 10 bits. More...
uint16_t   exponent_: 5
	The exponent composed of 5 bits. More...
uint16_t   sign_: 1
	The sign composed of 1 bit. More...
}   ieee_
	The struct with the IEEE representation. More...
}	data_
	Union allowing to access the binary representation or the IEEE representation. More...

Detailed Description

Definition of a 16-bit float type.

Constructor & Destructor Documentation

Float16() [1/3]

Ocean::Float16::Float16 ( )

default

Default constructor.

Float16() [2/3]

Ocean::Float16::Float16 ( const uint16_t  sign, const uint16_t  fraction, const uint16_t  exponent  )

inline

Creates a new 16-bit float.

Parameters

sign	The 1-bit sign to be used, with range [0, 1]
fraction	The 10-bit fraction to be used, with range [0, 1023]
exponent	The 5-bit exponent to be used, with range [0, 31]

Float16() [3/3]

Ocean::Float16::Float16 ( const uint16_t  binary )

inlineexplicit

Creates a new 16-bit float based on a 16-bit binary value.

Parameters

binary

The binary value to be used

Member Function Documentation

binary()

uint16_t Ocean::Float16::binary ( ) const

inline

Returns the binary representation of this 16-bit float value.

Returns

The value's binary representation

infinity()

Float16 Ocean::Float16::infinity ( )

inlinestatic

Returns +infinity.

Returns

The 16-bit float representing positive infinity

operator float()

Ocean::Float16::operator float ( ) const

explicit

Cast operator to 32-bit float.

Returns

The 32-bit float representation

operator!=()

bool Ocean::Float16::operator!= ( const Float16 &  second ) const

inline

Returns whether two 16-bit float values are not identical.

Parameters

second

The second value to compare

Returns

True, if so

operator-()

Float16 Ocean::Float16::operator- ( ) const

inline

Negates the 16-bit float value.

Returns

The negated value

operator==()

bool Ocean::Float16::operator== ( const Float16 &  second ) const

inline

Returns whether two 16-bit float values are identical.

Parameters

second

The second value to compare

Returns

True, if so

Field Documentation

binary_

uint16_t Ocean::Float16::binary_ = 0u

The binary representation.

union { ... } Ocean::Float16::data_

Union allowing to access the binary representation or the IEEE representation.

exponent_

uint16_t Ocean::Float16::exponent_

The exponent composed of 5 bits.

fraction_

uint16_t Ocean::Float16::fraction_

The mantissa composed of 10 bits.

struct { ... } Ocean::Float16::ieee_

The struct with the IEEE representation.

sign_

uint16_t Ocean::Float16::sign_

The sign composed of 1 bit.

---

The documentation for this class was generated from the following file:

• DataType.h