BFP API quick reference

The tables below list the functions of the block floating-point API. The “EW” column indicates whether the operation acts element-wise.

The “Signature” column is intended as a hint which quickly conveys the kind of the conceptual inputs to and outputs from the operation. The signatures are only intended to convey how many (conceptual) inputs and outputs there are, and their dimensionality.

The functions themselves will typically take more arguments than these signatures indicate. Check the function’s full documentation to get more detailed information.

The following symbols are used in the signatures:

Symbol	Description
\(\mathbb{S}\)	A scalar input or output value.
\(\mathbb{V}\)	A vector-valued input or output.
\(\mathbb{M}\)	A matrix-valued input or output.
\(\varnothing\)	Placeholder indicating no input or output.

For example, the operation signature \((\mathbb{V \times V \times S}) \to \mathbb{V}\) indicates the operation takes two vector inputs and a scalar input, and the output is a vector.

• 32-Bit BFP Ops

• 16-Bit BFP Ops

• Complex 32-Bit BFP Ops

• Complex 16-Bit BFP Ops

32-Bit BFP API quick reference

Table 3 32-Bit BFP API - quick reference

Function	EW	Signature	Brief
bfp_s32_init()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Initialize (static)
bfp_s32_alloc()		\(\varnothing \to \mathbb{V}\)	Initialize (dynamic)
bfp_s32_dealloc()		\(\mathbb{V} \to \mathbb{\varnothing}\)	Deinitialize
bfp_s32_set()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Set All Elements
bfp_s32_use_exponent()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Force Exponent
bfp_s32_headroom()		\(\mathbb{V} \to \mathbb{S}\)	Get Headroom
bfp_s32_shl()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Shift Mantissas
bfp_s32_add()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Add Vector
bfp_s32_add_scalar()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Add Scalar
bfp_s32_sub()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Subtract Vector
bfp_s32_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Multiply Vector
bfp_s32_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Multiply-Accumulate
bfp_s32_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Negated Multiply-Accumulate
bfp_s32_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Multiply Scalar
bfp_s32_abs()	x	\(\mathbb{V} \to \mathbb{V}\)	Absolute Values
bfp_s32_sum()		\(\mathbb{V} \to \mathbb{S}\)	Sum Elements
bfp_s32_dot()		\((\mathbb{V \times V}) \to \mathbb{S}\)	Inner Product
bfp_s32_clip()	x	\((\mathbb{V \times S \times S}) \to \mathbb{V}\)	Clip Bounds
bfp_s32_rect()	x	\(\mathbb{V} \to \mathbb{V}\)	Rectify Elements
bfp_s32_to_bfp_s16()		\(\mathbb{V} \to \mathbb{V}\)	Convert to 16-bit
bfp_s32_sqrt()	x	\(\mathbb{V} \to \mathbb{V}\)	Square Root
bfp_s32_inverse()	x	\(\mathbb{V} \to \mathbb{V}\)	Multiplicative Inverse
bfp_s32_abs_sum()		\(\mathbb{V} \to \mathbb{S}\)	Absolute Sum Elements
bfp_s32_mean()		\(\mathbb{V} \to \mathbb{S}\)	Vector Mean Value
bfp_s32_energy()		\(\mathbb{V} \to \mathbb{S}\)	Vector Energy
bfp_s32_rms()		\(\mathbb{V} \to \mathbb{S}\)	Vector RMS Value
bfp_s32_max()		\(\mathbb{V} \to \mathbb{S}\)	Vector Max Element
bfp_s32_min()		\(\mathbb{V} \to \mathbb{S}\)	Vector Min Element
bfp_s32_max_elementwise()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Elementwise Max
bfp_s32_min_elementwise()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Elementwise Min
bfp_s32_argmax()		\(\mathbb{V} \to \mathbb{S}\)	Max Element Index
bfp_s32_argmin()		\(\mathbb{V} \to \mathbb{S}\)	Min Element Index
bfp_s32_convolve_valid()		\((\mathbb{V \times V}) \to \mathbb{V}\)	Convolve With Kernel (Valid mode)
bfp_s32_convolve_same()		\((\mathbb{V \times V}) \to \mathbb{V}\)	Convolve With Kernel (Same mode)

16-Bit BFP API quick reference

Table 4 16-Bit BFP API - quick reference

Function	EW	Signature	Brief
bfp_s16_init()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Initialize (static)
bfp_s16_alloc()		\(\varnothing \to \mathbb{V}\)	Initialize (dynamic)
bfp_s16_dealloc()		\(\mathbb{V} \to \mathbb{\varnothing}\)	Deinitialize
bfp_s16_set()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Set All Elements
bfp_s16_use_exponent()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Force Exponent
bfp_s16_headroom()		\(\mathbb{V} \to \mathbb{S}\)	Get Headroom
bfp_s16_shl()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Shift Mantissas
bfp_s16_add()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Add Vector
bfp_s16_add_scalar()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Add Scalar
bfp_s16_sub()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Subtract Vector
bfp_s16_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Multiply Vector
bfp_s16_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Multiply-Accumulate
bfp_s16_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Negated Multiply-Accumulate
bfp_s16_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Multiply Scalar
bfp_s16_abs()	x	\(\mathbb{V} \to \mathbb{V}\)	Absolute Values
bfp_s16_sum()		\(\mathbb{V} \to \mathbb{S}\)	Sum Elements
bfp_s16_dot()		\((\mathbb{V \times V}) \to \mathbb{S}\)	Inner Product
bfp_s16_clip()	x	\((\mathbb{V \times S \times S}) \to \mathbb{V}\)	Clip Bounds
bfp_s16_rect()	x	\(\mathbb{V} \to \mathbb{V}\)	Rectify Elements
bfp_s16_to_bfp_s32()	x	\(\mathbb{V} \to \mathbb{V}\)	Convert to 32-bit
bfp_s16_sqrt()	x	\(\mathbb{V} \to \mathbb{V}\)	Square Root
bfp_s16_inverse()	x	\(\mathbb{V} \to \mathbb{V}\)	Multiplicative Inverse
bfp_s16_abs_sum()		\(\mathbb{V} \to \mathbb{V}\)	Absolute Sum Elements
bfp_s16_mean()		\(\mathbb{V} \to \mathbb{V}\)	Vector Mean Value
bfp_s16_energy()		\(\mathbb{V} \to \mathbb{S}\)	Vector Energy
bfp_s16_rms()		\(\mathbb{V} \to \mathbb{S}\)	Vector RMS Value
bfp_s16_max()		\(\mathbb{V} \to \mathbb{S}\)	Vector Max Element
bfp_s16_min()		\(\mathbb{V} \to \mathbb{S}\)	Vector Min Element
bfp_s16_max_elementwise()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Elementwise Max
bfp_s16_min_elementwise()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Elementwise Min
bfp_s16_argmax()		\(\mathbb{V} \to \mathbb{S}\)	Max Element Index
bfp_s16_argmin()		\(\mathbb{V} \to \mathbb{S}\)	Min Element Index
bfp_s16_accumulate()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Elementwise Accumulate

Complex 32-bit BFP API quick reference

Table 5 Complex 32-Bit BFP API - quick reference

Function	EW	Signature	Brief
bfp_complex_s32_init()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Initialize (static)
bfp_complex_s32_alloc()		\(\varnothing \to \mathbb{V}\)	Initialize (dynamic)
bfp_complex_s32_dealloc()		\(\mathbb{V} \to \mathbb{\varnothing}\)	Deinitialize
bfp_complex_s32_set()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Set All Elements
bfp_complex_s32_use_exponent()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Force Exponent
bfp_complex_s32_headroom()		\(\mathbb{V} \to \mathbb{S}\)	Get Headroom
bfp_complex_s32_shl()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Shift Mantissas
bfp_complex_s32_real_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Real Vector Multiply
bfp_complex_s32_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Multiply
bfp_complex_s32_conj_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Conjugate Multiply
bfp_complex_s32_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Multiply-Accumulate
bfp_complex_s32_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Negated Multiply-Accumulate
bfp_complex_s32_conj_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Conjugate Multiply-Accumulate
bfp_complex_s32_conj_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Negated Conjugate Multiply-Accumulate
bfp_complex_s32_real_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Real Scalar Multiply
bfp_complex_s32_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Complex Scalar Multiply
bfp_complex_s32_add()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Add
bfp_complex_s32_add_scalar()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Complex Scalar Add
bfp_complex_s32_sub()		\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Subtract
bfp_complex_s32_to_bfp_complex_s16()	x	\(\mathbb{V} \to \mathbb{V}\)	Convert to 16-bit
bfp_complex_s32_squared_mag()	x	\(\mathbb{V} \to \mathbb{V}\)	Squared Magnitude
bfp_complex_s32_mag()	x	\(\mathbb{V} \to \mathbb{V}\)	Magnitude
bfp_complex_s32_sum()		\(\mathbb{V} \to \mathbb{S}\)	Vector Sum
bfp_complex_s32_conjugate()	x	\(\mathbb{V} \to \mathbb{V}\)	Complex Conjugate
bfp_complex_s32_energy()		\(\mathbb{V} \to \mathbb{S}\)	Vector Energy
bfp_complex_s32_make()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Construct Complex From Real and Imaginary
bfp_complex_s32_real_part()	x	\(\mathbb{V} \to \mathbb{V}\)	Real Part
bfp_complex_s32_imag_part()	x	\(\mathbb{V} \to \mathbb{V}\)	Imaginary Part

Complex 16-bit BFP API quick reference

Table 6 Complex 16-Bit BFP API - quick reference

Function	EW	Signature	Brief
bfp_complex_s16_init()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Initialize (static)
bfp_complex_s16_alloc()		\(\varnothing \to \mathbb{V}\)	Initialize (dynamic)
bfp_complex_s16_dealloc()		\(\mathbb{V} \to \mathbb{\varnothing}\)	Deinitialize
bfp_complex_s16_set()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Set All Elements
bfp_complex_s16_use_exponent()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Force Exponent
bfp_complex_s16_headroom()		\(\mathbb{V} \to \mathbb{S}\)	Get Headroom
bfp_complex_s16_shl()	x	\((\mathbb{V \times S}) \to \mathbb{V}\)	Shift Mantissas
bfp_complex_s16_real_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Real Vector Multiply
bfp_complex_s16_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Multiply
bfp_complex_s16_conj_mul()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Conjugate Multiply
bfp_complex_s16_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Multiply-Accumulate
bfp_complex_s16_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Negated Multiply-Accumulate
bfp_complex_s16_conj_macc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Conjugate Multiply-Accumulate
bfp_complex_s16_conj_nmacc()	x	\((\mathbb{V \times V \times V}) \to \mathbb{V}\)	Complex Vector Negated Conjugate Multiply-Accumulate
bfp_complex_s16_real_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Real Scalar Multiply
bfp_complex_s16_scale()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Complex Scalar Multiply
bfp_complex_s16_add()	x	\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Add
bfp_complex_s16_add_scalar()		\((\mathbb{V \times S}) \to \mathbb{V}\)	Complex Scalar Add
bfp_complex_s16_sub()		\((\mathbb{V \times V}) \to \mathbb{V}\)	Complex Vector Subtract
bfp_complex_s16_to_bfp_complex_s32()	x	\(\mathbb{V} \to \mathbb{V}\)	Convert to 32-bit
bfp_complex_s16_squared_mag()	x	\(\mathbb{V} \to \mathbb{V}\)	Squared Magnitude
bfp_complex_s16_sum()		\(\mathbb{V} \to \mathbb{S}\)	Vector Sum
bfp_complex_s16_mag()	x	\(\mathbb{V} \to \mathbb{V}\)	Magnitude
bfp_complex_s16_conjugate()	x	\(\mathbb{V} \to \mathbb{V}\)	Complex Conjugate
bfp_complex_s16_energy()		\(\mathbb{V} \to \mathbb{S}\)	Vector Energy