Halide 19.0.0
Halide compiler and libraries
Loading...
Searching...
No Matches
Halide::float16_t Struct Reference

Class that provides a type that implements half precision floating point (IEEE754 2008 binary16) in software. More...

#include <Float16.h>

Public Member Functions

 operator float () const
 Cast to float.
 
 operator double () const
 Cast to double.
 
 operator int () const
 Cast to int.
 
float16_t operator- () const
 Return a new float16_t with a negated sign bit.
 
float16_t operator+ (float16_t rhs) const
 Arithmetic operators.
 
float16_t operator- (float16_t rhs) const
 
float16_t operator* (float16_t rhs) const
 
float16_t operator/ (float16_t rhs) const
 
float16_t operator+= (float16_t rhs)
 
float16_t operator-= (float16_t rhs)
 
float16_t operator*= (float16_t rhs)
 
float16_t operator/= (float16_t rhs)
 
bool operator== (float16_t rhs) const
 Comparison operators.
 
bool operator!= (float16_t rhs) const
 
bool operator> (float16_t rhs) const
 
bool operator< (float16_t rhs) const
 
bool operator>= (float16_t rhs) const
 
bool operator<= (float16_t rhs) const
 
bool is_nan () const
 Properties.
 
bool is_infinity () const
 
bool is_negative () const
 
bool is_zero () const
 
uint16_t to_bits () const
 Returns the bits that represent this float16_t.
 
Constructors
 float16_t (float value)
 Construct from a float, double, or int using round-to-nearest-ties-to-even.
 
 float16_t (double value)
 Construct from a float, double, or int using round-to-nearest-ties-to-even.
 
 float16_t (int value)
 Construct from a float, double, or int using round-to-nearest-ties-to-even.
 
 float16_t ()=default
 Construct a float16_t with the bits initialised to 0.
 

Static Public Member Functions

static float16_t make_zero ()
 Get a new float16_t that represents a special value.
 
static float16_t make_negative_zero ()
 
static float16_t make_infinity ()
 
static float16_t make_negative_infinity ()
 
static float16_t make_nan ()
 
static float16_t make_from_bits (uint16_t bits)
 Get a new float16_t with the given raw bits.
 

Static Public Attributes

static const int mantissa_bits = 10
 
static const uint16_t sign_mask = 0x8000
 
static const uint16_t exponent_mask = 0x7c00
 
static const uint16_t mantissa_mask = 0x03ff
 

Detailed Description

Class that provides a type that implements half precision floating point (IEEE754 2008 binary16) in software.

This type is enforced to be 16-bits wide and maintains no state other than the raw IEEE754 binary16 bits so that it can passed to code that checks a type's size and used for halide_buffer_t allocation.

Definition at line 17 of file Float16.h.

Constructor & Destructor Documentation

◆ float16_t() [1/4]

Halide::float16_t::float16_t ( float value)
explicit

Construct from a float, double, or int using round-to-nearest-ties-to-even.

Out-of-range values become +/- infinity.

◆ float16_t() [2/4]

Halide::float16_t::float16_t ( double value)
explicit

Construct from a float, double, or int using round-to-nearest-ties-to-even.

Out-of-range values become +/- infinity.

◆ float16_t() [3/4]

Halide::float16_t::float16_t ( int value)
explicit

Construct from a float, double, or int using round-to-nearest-ties-to-even.

Out-of-range values become +/- infinity.

◆ float16_t() [4/4]

Halide::float16_t::float16_t ( )
default

Construct a float16_t with the bits initialised to 0.

This represents positive zero.

References memcpy().

Member Function Documentation

◆ operator float()

Halide::float16_t::operator float ( ) const
explicit

Cast to float.

◆ operator double()

Halide::float16_t::operator double ( ) const
explicit

Cast to double.

◆ operator int()

Halide::float16_t::operator int ( ) const
explicit

Cast to int.

◆ make_zero()

static float16_t Halide::float16_t::make_zero ( )
static

Get a new float16_t that represents a special value.

◆ make_negative_zero()

static float16_t Halide::float16_t::make_negative_zero ( )
static

◆ make_infinity()

static float16_t Halide::float16_t::make_infinity ( )
static

◆ make_negative_infinity()

static float16_t Halide::float16_t::make_negative_infinity ( )
static

◆ make_nan()

static float16_t Halide::float16_t::make_nan ( )
static

◆ make_from_bits()

static float16_t Halide::float16_t::make_from_bits ( uint16_t bits)
static

Get a new float16_t with the given raw bits.

Parameters
bitsThe bits conformant to IEEE754 binary16

◆ operator-() [1/2]

float16_t Halide::float16_t::operator- ( ) const

Return a new float16_t with a negated sign bit.

◆ operator+()

float16_t Halide::float16_t::operator+ ( float16_t rhs) const

Arithmetic operators.

◆ operator-() [2/2]

float16_t Halide::float16_t::operator- ( float16_t rhs) const

◆ operator*()

float16_t Halide::float16_t::operator* ( float16_t rhs) const

◆ operator/()

float16_t Halide::float16_t::operator/ ( float16_t rhs) const

◆ operator+=()

float16_t Halide::float16_t::operator+= ( float16_t rhs)
inline

Definition at line 91 of file Float16.h.

◆ operator-=()

float16_t Halide::float16_t::operator-= ( float16_t rhs)
inline

Definition at line 94 of file Float16.h.

◆ operator*=()

float16_t Halide::float16_t::operator*= ( float16_t rhs)
inline

Definition at line 97 of file Float16.h.

◆ operator/=()

float16_t Halide::float16_t::operator/= ( float16_t rhs)
inline

Definition at line 100 of file Float16.h.

◆ operator==()

bool Halide::float16_t::operator== ( float16_t rhs) const

Comparison operators.

◆ operator!=()

bool Halide::float16_t::operator!= ( float16_t rhs) const
inline

Definition at line 108 of file Float16.h.

◆ operator>()

bool Halide::float16_t::operator> ( float16_t rhs) const

◆ operator<()

bool Halide::float16_t::operator< ( float16_t rhs) const

◆ operator>=()

bool Halide::float16_t::operator>= ( float16_t rhs) const
inline

Definition at line 113 of file Float16.h.

◆ operator<=()

bool Halide::float16_t::operator<= ( float16_t rhs) const
inline

Definition at line 116 of file Float16.h.

◆ is_nan()

bool Halide::float16_t::is_nan ( ) const

Properties.

◆ is_infinity()

bool Halide::float16_t::is_infinity ( ) const

◆ is_negative()

bool Halide::float16_t::is_negative ( ) const

◆ is_zero()

bool Halide::float16_t::is_zero ( ) const

◆ to_bits()

uint16_t Halide::float16_t::to_bits ( ) const

Returns the bits that represent this float16_t.

An alternative method to access the bits is to cast a pointer to this instance as a pointer to a uint16_t.

Member Data Documentation

◆ mantissa_bits

const int Halide::float16_t::mantissa_bits = 10
static

Definition at line 19 of file Float16.h.

◆ sign_mask

const uint16_t Halide::float16_t::sign_mask = 0x8000
static

Definition at line 20 of file Float16.h.

◆ exponent_mask

const uint16_t Halide::float16_t::exponent_mask = 0x7c00
static

Definition at line 21 of file Float16.h.

◆ mantissa_mask

const uint16_t Halide::float16_t::mantissa_mask = 0x03ff
static

Definition at line 22 of file Float16.h.


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