#include <binary32.hh>

Public Types
enum	bitSizes { ebits = 8 , mbits = 23 , sbits = 1 , bias = 127 , inf = 0x7f800000 , nan = 0x7f800100 , max = 0x7f7fffff }

Public Member Functions
	operator float () const

Public Attributes
uint32_t	storage
float	fp32
struct {
unsigned mant: 23
unsigned exp: 8
unsigned sign: 1
};

Detailed Description

Definition at line 45 of file binary32.hh.

Member Enumeration Documentation

◆ bitSizes

enum gem5::AMDGPU::binary32_u::bitSizes

Enumerator
ebits
mbits
sbits
bias
inf
nan
max

Definition at line 47 of file binary32.hh.

Member Function Documentation

◆ operator float()

gem5::AMDGPU::binary32_u::operator float ( ) const

inline

Definition at line 69 of file binary32.hh.

References fp32.

Member Data Documentation

◆ [struct]

struct { ... } gem5::AMDGPU::binary32_u

◆ exp

unsigned gem5::AMDGPU::binary32_u::exp

Definition at line 64 of file binary32.hh.

Referenced by gem5::AMDGPU::mxfp< fp8_e5m2_info >::scaleDiv(), and gem5::AMDGPU::mxfp< fp8_e5m2_info >::scaleMul().

◆ fp32

float gem5::AMDGPU::binary32_u::fp32

Definition at line 60 of file binary32.hh.

Referenced by gem5::AMDGPU::mxfp< fp8_e5m2_info >::float_to_mxfp(), std::numeric_limits< gem5::AMDGPU::binary32 >::infinity(), std::numeric_limits< gem5::AMDGPU::binary32 >::max(), operator float(), gem5::AMDGPU::mxfp< fp8_e5m2_info >::operator float(), std::numeric_limits< gem5::AMDGPU::binary32 >::quiet_NaN(), gem5::AMDGPU::mxfp< fp8_e5m2_info >::scaleDiv(), and gem5::AMDGPU::mxfp< fp8_e5m2_info >::scaleMul().

◆ mant

unsigned gem5::AMDGPU::binary32_u::mant

Definition at line 63 of file binary32.hh.

◆ sign

unsigned gem5::AMDGPU::binary32_u::sign

Definition at line 65 of file binary32.hh.

◆ storage

uint32_t gem5::AMDGPU::binary32_u::storage

Definition at line 59 of file binary32.hh.

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

arch/amdgpu/common/dtype/binary32.hh

Public Types

Public Member Functions

Public Attributes