Floating-point Conversion

Overview

Supported float to fixed conversions

			bfloat16		float
Output bits	Type	Arch.	Implementation	Notes	Implementation	Notes
4b	Scalar	AIE
		AIE-ML	Runs on vector unit		Runs on vector unit
	Vector	AIE
		AIE-ML	Emulated	Uses aie::rounding_mode::symmetric_zero	Emulated	Uses aie::rounding_mode::symmetric_zero
8b	Scalar	AIE			Native to 32b + cast
		AIE-ML	Runs on vector unit		Runs on vector unit
	Vector	AIE			Element-wise scalar
		AIE-ML	Emulated	Uses aie::rounding_mode::symmetric_zero	Emulated	Uses aie::rounding_mode::symmetric_zero
16b	Scalar	AIE			Native to 32b + cast
		AIE-ML	Runs on vector unit		Runs on vector unit
	Vector	AIE			Vectorized emulated impl
		AIE-ML	Native to 32b + extract lower 16b1	Uses aie::rounding_mode::floor	Emulated	Uses aie::rounding_mode::symmetric_zero
32b	Scalar	AIE			Native
		AIE-ML	Runs on vector unit		Runs on vector unit
	Vector	AIE			Vectorized emulated impl
		AIE-ML	Native	Uses aie::rounding_mode::floor	Emulated	Uses aie::rounding_mode::symmetric_zero

Note

¹ - Unsigned conversions are emulated and use aie::rounding_mode::symmetric_zero.

Supported fixed to float conversions

Output type	Type	Arch.	int4	int8	int16	int32
bfloat16	Scalar	AIE
		AIE-ML	Runs on vector unit	Runs on vector unit	Runs on vector unit	Runs on vector unit
	Vector	AIE
		AIE-ML	Emulated	Emulated	Emulated	Emulated
float	Scalar	AIE		Cast to int32 + native	Cast to int32 + native	Native
		AIE-ML	Runs on vector unit	Runs on vector unit	Runs on vector unit	Runs on vector unit
	Vector	AIE		Element-wise scalar	Emulated	Element-wise scalar
		AIE-ML	Emulated	Emulated	Emulated	Emulated

Functions
template<typename TR = cint32> requires (arch::is(arch::AIE, arch::AIE_ML) && Utils::is_one_of_v<TR, cint16, cint32>)
auto	aie::to_fixed (cfloat a, int shift=0)
	Convert a floating point value into a fixed-point value.
template<typename TR = int32, unsigned Elems> requires (arch::is(arch::AIE_ML))
auto	aie::to_fixed (const vector< bfloat16, Elems > &v, int shift=0) -> vector< TR, Elems >
	Convert the elements in a bfloat16 vector into fixed-point values.
template<typename TR = cint32, unsigned Elems> requires ((arch::is(arch::AIE) || (arch::is(arch::AIE_ML) && __AIE_API_COMPLEX_FP32_EMULATION__ == 1)) && Utils::is_one_of_v<TR, cint16, cint32>)
auto	aie::to_fixed (const vector< cfloat, Elems > &v, int shift=0) -> vector< TR, Elems >
	Convert the elements in a floating point vector into fixed-point values.
template<typename TR = int32, unsigned Elems> requires (Utils::is_one_of_v<TR, int16, int32>)
auto	aie::to_fixed (const vector< float, Elems > &v, int shift=0) -> vector< TR, Elems >
	Convert the elements in a floating point vector into fixed-point values.
template<typename TR = int32> requires (Utils::is_one_of_v<TR, int8, int16, int32>)
auto	aie::to_fixed (float a, int shift=0)
	Convert a floating point value into a fixed-point value.
template<typename TR = float, typename T , unsigned Elems> requires (Utils::is_one_of_v<T, int16, int32>)
auto	aie::to_float (const vector< T, Elems > &v, int shift=0) -> vector< TR, Elems >
	Convert the elements in a real fixed-point vector into floating point values.
template<typename TR = float>
auto	aie::to_float (int a, int shift=0)
	Convert a fixed-point value into a floating point value.
template<typename TR = cfloat, typename T > requires (Utils::is_one_of_v<T, cint16, cint32>)
auto	aie::to_float (T a, int shift=0)
	Convert a complex fixed-point value into a floating point value.

Function Documentation

to_fixed() [1/5]

template<typename TR = cint32>
requires (arch::is(arch::AIE, arch::AIE_ML) && Utils::is_one_of_v<TR, cint16, cint32>)

auto aie::to_fixed	(	cfloat	a,
		int	shift = 0
	)

Convert a floating point value into a fixed-point value.

Parameters

a	Input value.
shift	Position of the point in the output value.

Template Parameters

TR	Type of the returned fixed-point value.

to_fixed() [2/5]

template<typename TR = int32, unsigned Elems>
requires (arch::is(arch::AIE_ML))

auto aie::to_fixed	(	const vector< bfloat16, Elems > &	v,
		int	shift = 0
	)		-> vector<TR, Elems>

Convert the elements in a bfloat16 vector into fixed-point values.

Note

This conversion performs floor rounding, regardless of the rounding_mode configured by the user.

Parameters

v	Input vector.
shift	Position of the point in the output values.

Template Parameters

TR	Type of the returned fixed-point values.

to_fixed() [3/5]

template<typename TR = cint32, unsigned Elems>
requires ((arch::is(arch::AIE) || (arch::is(arch::AIE_ML) && __AIE_API_COMPLEX_FP32_EMULATION__ == 1)) && Utils::is_one_of_v<TR, cint16, cint32>)

auto aie::to_fixed	(	const vector< cfloat, Elems > &	v,
		int	shift = 0
	)		-> vector<TR, Elems>

Convert the elements in a floating point vector into fixed-point values.

Parameters

v	Input vector.
shift	Position of the point in the output values.

Template Parameters

TR	Type of the returned fixed-point values.

to_fixed() [4/5]

template<typename TR = int32, unsigned Elems>
requires (Utils::is_one_of_v<TR, int16, int32>)

auto aie::to_fixed	(	const vector< float, Elems > &	v,
		int	shift = 0
	)		-> vector<TR, Elems>

Convert the elements in a floating point vector into fixed-point values.

Parameters

v	Input vector.
shift	Position of the point in the output values.

Template Parameters

TR	Type of the returned fixed-point values.

to_fixed() [5/5]

template<typename TR = int32>
requires (Utils::is_one_of_v<TR, int8, int16, int32>)

auto aie::to_fixed	(	float	a,
		int	shift = 0
	)

Convert a floating point value into a fixed-point value.

Parameters

a	Input value.
shift	Position of the point in the output value.

Template Parameters

TR	Type of the returned fixed-point value.

to_float() [1/3]

template<typename TR = float, typename T , unsigned Elems>
requires (Utils::is_one_of_v<T, int16, int32>)

auto aie::to_float	(	const vector< T, Elems > &	v,
		int	shift = 0
	)		-> vector<TR, Elems>

Convert the elements in a real fixed-point vector into floating point values.

Parameters

v	Input vector.
shift	Position of the point in the input values.

to_float() [2/3]

template<typename TR = float>

auto aie::to_float	(	int	a,
		int	shift = 0
	)

Convert a fixed-point value into a floating point value.

Parameters

a	Input value.
shift	Position of the point in the input value.

to_float() [3/3]

template<typename TR = cfloat, typename T >
requires (Utils::is_one_of_v<T, cint16, cint32>)

auto aie::to_float	(	T	a,
		int	shift = 0
	)

Convert a complex fixed-point value into a floating point value.

Parameters

a	Input value.
shift	Position of the point in the input value.