ieee754.h File Reference

IEEE754 floating point manipulation. More...

#include <wfio/type.h>
#include <wfio/byteswap.h>
#include <wfcnfo/float.h>

Go to the source code of this file.

Classes
union	wfio_ieee754_b16_t
	A half-precision IEEE 754 binary floating point value (binary16). More...
union	wfio_ieee754_b32_t
	A single-precision IEEE 754 binary floating point value (binary32). More...
union	wfio_ieee754_b64_t
	A double-precision IEEE 754 binary floating point value (binary64). More...

Macros
#define	WFIO_IEEE754_NATIVE_B16
	A macro indicating that there is native support for IEEE 754 numbers in binary16 format.
#define	WFIO_IEEE754_NATIVE_B32
	A macro indicating that there is native support for IEEE 754 numbers in binary32 format.
#define	WFIO_IEEE754_NATIVE_B64
	A macro indicating that there is native support for IEEE 754 numbers in binary64 format.
#define	WFIO_IEEE754_NATIVE
	Synonymous with WFIO_IEEE754_NATIVE_B32 && WFIO_IEEE_NATIVE_B64.
#define	WFIO_IEEE754_B32_MIN   (wfio_ieee854_b32_t) 0x1
	A constant representing the minimum positive, nonzero value for an IEEE 754 binary32 number.
#define	WFIO_IEEE754_B32_MAX   (wfio_ieee754_b32_t) 0x7EFFFFFF
	A constant representing the maximum positive, non-infinite value for an IEEE 754 binary32 number.
#define	WFIO_IEEE754_B32_INF   (wfio_ieee754_b32_t) 0x7F000000
	A constant representing the positive infinite value for an IEEE 754 binary32 number.
#define	WFIO_IEEE754_B32_ENAN   (wfio_ieee754_b32_t) 0x7FFFFFFF

Typedefs
typedef WFIO_C_BEGIN union wfio_ieee754_b16_t	wfio_ieee754_b16_t
	A half-precision IEEE 754 binary floating point value (binary16). More...
typedef union wfio_ieee754_b32_t	wfio_ieee754_b32_t
	A single-precision IEEE 754 binary floating point value (binary32). More...
typedef union wfio_ieee754_b64_t	wfio_ieee754_b64_t
	A double-precision IEEE 754 binary floating point value (binary64). More...

Enumerations
enum	{ WFIO_IEEE754_B32_ENAN_RAW = 0x7FFFFFFF, WFIO_IEEE754_B64_ENAN_RAW = 0x7FFFFFFFFFFFFFFF }

Functions
WFIO_MAYBE_INLINE wfio_ieee754_b32_t	wfio_ieee754_toB32 (wfio_flt32_t host, int *error)
	Convert a 32-bit floating point number to IEEE 754 binary32 format. More...
WFIO_MAYBE_INLINE wfio_flt32_t	wfio_ieee754_fromB32 (wfio_ieee754_b32_t ieee, int *error)
	Convert a 32-bit floating point number from IEEE 754 binary32 format. More...
WFIO_DLL size_t	wfio_ieee754_toB32V2 (const wfio_flt32_t *src, wfio_ieee754_b32_t *dest, size_t len, int *error)
WFIO_DLL size_t	wfio_ieee754_fromB32V2 (const wfio_ieee754_b32_t *src, wfio_flt32_t *dest, size_t len, int *error)
WFIO_DLL wfio_ieee754_b64_t	wfio_ieee754_toB64 (wfio_flt64_t host, int *error)
	Convert a 64-bit floating point number to IEEE 754 binary64 format. More...
WFIO_MAYBE_INLINE wfio_flt64_t	wfio_ieee754_fromB64 (wfio_ieee754_b64_t ieee, int *error)
	Convert a 64-bit floating point number from IEEE 754 binary64 format. More...
WFIO_DLL size_t	wfio_ieee754_toB64V2 (const wfio_flt64_t *src, wfio_ieee754_b64_t *dest, size_t len, int *error)
WFIO_DLL size_t	wfio_ieee754_fromB64V2 (const wfio_ieee754_b64_t *src, wfio_flt64_t *dest, size_t len, int *error)

Detailed Description

IEEE754 floating point manipulation.

This header provides IEEE 754 support for architectures which do not support it. This is necessary because although most modern machines do implement this standard, many do not. Furthermore, there have been extensions to the standard since when it was first published in 1985. This header is provided so that these numbers may be easily written to a device or stream.

This header only implements the basic binary formats. For extended support, link against libwfion and include wfio/num/ieee754.h.

Author

Phillip Kilgore

Since

0.0.1

Typedef Documentation

wfio_ieee754_b16_t

typedef WFIO_C_BEGIN union wfio_ieee754_b16_t wfio_ieee754_b16_t

A half-precision IEEE 754 binary floating point value (binary16).

This type arises from IEEE 754-2008 and represents a half-precision floating point number. This can fit within exactly 16 bits and may be used in situations where the range does not exceed ± ~2^15. in total, there are 11 bits of precision in the significand.

wfio_ieee754_b32_t

typedef union wfio_ieee754_b32_t wfio_ieee754_b32_t

A single-precision IEEE 754 binary floating point value (binary32).

This type arises from IEEE 754-1985 and represents a single-precision floating point number. This can fit within exactly 32 bits and may be used in situations where the range does not exceed ± ~2^127. In total, there are 24 bits of precision in the significand.

wfio_ieee754_b64_t

typedef union wfio_ieee754_b64_t wfio_ieee754_b64_t

A double-precision IEEE 754 binary floating point value (binary64).

This type arises from IEEE 754-1985 and represents a single-precision floating point number. This can fit within exactly 32 bits and may be used in situations where the range does not exceed ± ~2^1024. In total, there are 53 bits of precision in the significand.

Function Documentation

wfio_ieee754_fromB32()

WFIO_MAYBE_INLINE wfio_flt32_t wfio_ieee754_fromB32	(	wfio_ieee754_b32_t	ieee,
		int *	error
	)

Convert a 32-bit floating point number from IEEE 754 binary32 format.

Parameters

ieee	The number to convert.

Returns

A value in the host floating point format.

wfio_ieee754_fromB64()

WFIO_MAYBE_INLINE wfio_flt64_t wfio_ieee754_fromB64	(	wfio_ieee754_b64_t	ieee,
		int *	error
	)

Convert a 64-bit floating point number from IEEE 754 binary64 format.

Parameters

ieee	The number to convert.

Returns

A value in the host floating point format.

wfio_ieee754_toB32()

WFIO_MAYBE_INLINE wfio_ieee754_b32_t wfio_ieee754_toB32	(	wfio_flt32_t	host,
		int *	error
	)

Convert a 32-bit floating point number to IEEE 754 binary32 format.

Parameters

host	The number to convert.

Returns

A value in IEEE 754 binary32 format.

wfio_ieee754_toB64()

WFIO_DLL wfio_ieee754_b64_t wfio_ieee754_toB64	(	wfio_flt64_t	host,
		int *	error
	)

Convert a 64-bit floating point number to IEEE 754 binary64 format.

Parameters

host	The number to convert.

Returns

A value in IEEE 754 binary64 format.