Burger::Half Class Reference

Container for a 16 bit float. More...

Public Member Functions
	Half () noexcept
	Constructor.
	Half (float16_t uData) noexcept
	Constructor with 16 bit float.
	Half (float fInput) noexcept
	Constructor with 32 bit float.
	operator float () const noexcept
	Convert to a 32 bit float.
	operator uint16_t () const noexcept
	Return the 16 bit float.
uint_t	is_finite () const noexcept
	Test if the value is finite.
uint_t	is_infinity () const noexcept
	Test if the value is infinite.
uint_t	is_nan () const noexcept
	Test if the value is NaN.
uint_t	is_normalized () const noexcept
	Test if the value is normalzed.
uint_t	is_denormalized () const noexcept
	Test if the value is denormalzed.
uint_t	is_zero () const noexcept
	Test if the value is zero.
uint_t	is_negative () const noexcept
	Test if the value is negative.
Half	operator- () const noexcept
	Return a negated Half.
Half &	operator= (const Half &rInput) noexcept
	Copy a Half class.
Half &	operator= (float fInput) noexcept
	Assign a 32 bit float to this class.
Half &	operator+= (const Half &rInput) noexcept
	Add a Half to this Half.
Half &	operator+= (float fInput) noexcept
	Add a float to this Half.
Half &	operator-= (const Half &rInput) noexcept
	Subtract a Half from this Half.
Half &	operator-= (float fInput) noexcept
	Subtract a float to this Half.
Half &	operator*= (const Half &rInput) noexcept
	Multiply a Half with this Half.
Half &	operator*= (float fInput) noexcept
	Multiply a float to this Half.
Half &	operator/= (const Half &rInput) noexcept
	Divide a Half from this Half.
Half &	operator/= (float fInput) noexcept
	Divide a float from this Half.

Private Attributes
float16_t	m_uData
	Binary representation of a 16 bit float.

Detailed Description

Container for a 16 bit float.

---

This class will allow math functions to be applied to a 16 bit float since most "C" compilers don't have native support for the data type.

While this class exists, it is recommended to perform math operations with float or double until a final result is created and then the result is converted to the float16 data type.

See also

convert_to_float16(float), or convert_to_float(float16_t)

Constructor & Destructor Documentation

Half() [1/3]

Burger::Half::Half ( )

inlinenoexcept

Constructor.

---

This constructor does NOT initialize this class.

See also

Half(float16_t), or Half(float)

Half() [2/3]

Burger::Half::Half ( float16_t uData )

inlinenoexcept

Constructor with 16 bit float.

---

Initialize with a 16 bit float.

See also

Half(float), or Half()

Half() [3/3]

Burger::Half::Half ( float fInput )

noexcept

Constructor with 32 bit float.

---

Initialize with a 32 bit float, which is converted into a 16 bit float.

See also

Half(float16_t), or Half()

Member Function Documentation

is_denormalized()

uint_t Burger::Half::is_denormalized ( ) const

inlinenoexcept

Test if the value is denormalzed.

---

If the contained value is denormalized, return TRUE

Returns

TRUE or FALSE

See also

is_normalized() const

is_finite()

uint_t Burger::Half::is_finite ( ) const

inlinenoexcept

Test if the value is finite.

---

If the contained value is a finite value, return TRUE

Returns

TRUE or FALSE

See also

is_infinity() const, or is_nan() const

is_infinity()

uint_t Burger::Half::is_infinity ( ) const

inlinenoexcept

Test if the value is infinite.

---

If the contained value is a infinite value, return TRUE

Returns

TRUE or FALSE

See also

is_finite() const, or is_nan() const

is_nan()

uint_t Burger::Half::is_nan ( ) const

inlinenoexcept

Test if the value is NaN.

---

If the contained value is a NaN (Not a Number) value, return TRUE

Returns

TRUE or FALSE

See also

is_finite() const, or is_infinity() const

is_negative()

uint_t Burger::Half::is_negative ( ) const

inlinenoexcept

Test if the value is negative.

---

If the contained value has the sign bit set, return TRUE

Returns

TRUE or FALSE

See also

is_zero() const

is_normalized()

uint_t Burger::Half::is_normalized ( ) const

inlinenoexcept

Test if the value is normalzed.

---

If the contained value is normalized, return TRUE

Returns

TRUE or FALSE

See also

is_denormalized() const

is_zero()

uint_t Burger::Half::is_zero ( ) const

inlinenoexcept

Test if the value is zero.

---

If the contained value is zero, return TRUE

Note

It will check for both positive and negative zero.

Returns

TRUE or FALSE

See also

is_negative() const

operator float()

Burger::Half::operator float ( ) const

noexcept

Convert to a 32 bit float.

---

Convert the float 16 into a 32 bit float and return the float.

See also

operator uint16_t()

operator uint16_t()

Burger::Half::operator uint16_t ( ) const

inlinenoexcept

Return the 16 bit float.

---

See also

operator float()

operator*=() [1/2]

Burger::Half & Burger::Half::operator*= ( const Half & rInput )

noexcept

Multiply a Half with this Half.

---

Convert both inputs into floats, perform a multiplication, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

rInput

Half to multiply to this instance.

Returns

*this

operator*=() [2/2]

Burger::Half & Burger::Half::operator*= ( float fInput )

noexcept

Multiply a float to this Half.

---

Convert this class into a float, perform a multiplication, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

fInput

32 bit float to multiply with this Half

Returns

*this

operator+=() [1/2]

Burger::Half & Burger::Half::operator+= ( const Half & rInput )

noexcept

Add a Half to this Half.

---

Convert both inputs into floats, perform an addition, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

rInput

Half to add to this instance.

Returns

*this

operator+=() [2/2]

Burger::Half & Burger::Half::operator+= ( float fInput )

noexcept

Add a float to this Half.

---

Convert this class into a float, perform an addition, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

fInput

32 bit float to add to this Half

Returns

*this

operator-()

Burger::Half Burger::Half::operator- ( ) const

noexcept

Return a negated Half.

---

Negate the float16 and return that result as a newly created Half class.

Returns

Half class with the value negated from this value

operator-=() [1/2]

Burger::Half & Burger::Half::operator-= ( const Half & rInput )

noexcept

Subtract a Half from this Half.

---

Convert both inputs into floats, perform a subtraction, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

rInput

Half to subtract from this instance.

Returns

*this

operator-=() [2/2]

Burger::Half & Burger::Half::operator-= ( float fInput )

noexcept

Subtract a float to this Half.

---

Convert this class into a float, perform a subtraction, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

fInput

32 bit float to subtract from this Half

Returns

*this

operator/=() [1/2]

Burger::Half & Burger::Half::operator/= ( const Half & rInput )

noexcept

Divide a Half from this Half.

---

Convert both inputs into floats, perform a division, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

rInput

Half to divide from this instance.

Returns

*this

operator/=() [2/2]

Burger::Half & Burger::Half::operator/= ( float fInput )

noexcept

Divide a float from this Half.

---

Convert this class into a float, perform a division, and then store the result in this class.

Note

Due to conversions, this function is not as fast as performing hardware float operations.

Parameters

fInput

32 bit float to divide from this Half

Returns

*this

operator=() [1/2]

Burger::Half & Burger::Half::operator= ( const Half & rInput )

noexcept

Copy a Half class.

---

Copy the contents of the supplied Half into this instance.

Parameters

rInput

Half to copy to this instance.

Returns

*this

operator=() [2/2]

Burger::Half & Burger::Half::operator= ( float fInput )

noexcept

Assign a 32 bit float to this class.

---

Convert the 32 bit float into a 16 bit float and store the 16 bit value in this instance of this class.

Parameters

fInput

32 bit float to convert to a 16 bit float.

Returns

*this

Member Data Documentation

m_uData

float16_t Burger::Half::m_uData

private

Binary representation of a 16 bit float.