Overview

The MCUXpresso SDK provides driver for the PowerQuad module of MCUXpresso SDK devices.

The PowerQuad hardware accelerator for (fixed/floating point/matrix operation) DSP functions is that idea is to replace some of the CMSIS DSP functionality with the hardware features provided by this IP.

PowerQuad driver provides the following CMSIS DSP compatible functions:

Matrix functions
arm_mat_add_q15

arm_mat_add_q31

arm_mat_add_f32

arm_mat_sub_q15

arm_mat_sub_q31

arm_mat_sub_f32

arm_mat_mult_q15

arm_mat_mult_q31

arm_mat_mult_f32

arm_mat_inverse_f32

arm_mat_trans_q15

arm_mat_trans_q31

arm_mat_trans_f32

arm_mat_scale_q15

arm_mat_scale_q31

arm_mat_scale_f32
Math functions
arm_sqrt_q15

arm_sqrt_q31

arm_sin_q15

arm_sin_q31

arm_sin_f32

arm_cos_q15

arm_cos_q31

arm_cos_f32
Filter functions
arm_fir_q15

arm_fir_q31

arm_fir_f32

arm_conv_q15

arm_conv_q31

arm_conv_f32

arm_correlate_q15

arm_correlate_q31

arm_correlate_f32
Transform functions
arm_rfft_q15

arm_rfft_q31

arm_cfft_q15

arm_cfft_q31

arm_ifft_q15

arm_ifft_q31

arm_dct4_q15

arm_dct4_q31

Note

CMSIS DSP compatible functions limitations

PowerQuad FFT engine only looks at the bottom 27 bits of the input word, down scale the input data to avoid saturation.
When use arm_fir_q15/arm_fir_q31/arm_fir_f32 for incremental, the new data should follow the old data. For example the array for input data is inputData[], and the array for output data is outputData[]. The first 32 input data is saved in inputData[0:31], after callling arm_fir_xxx(&fir, inputData, outputData, 32), the output data is saved in outputData[0:31]. The new input data must be saved from inputData[32], then call arm_fir_xxx(&fir, &inputData[32], &outputData[32], 32) for incremental calculation.

The PowerQuad consists of several internal computation engines: Transform engine, Transcendental function engine, Trigonometry function engine, Dual biquad IIR filter engine, Matrix accelerator engine, FIR filter engine, CORDIC engine.

For low level APIs, all function APIs, except using coprocessor instruction and arctan/arctanh API, need to calling wait done API to wait for calculation complete.

Function groups

POWERQUAD functional Operation

This group implements the POWERQUAD functional API.

Data Structures
struct	pq_prescale_t
	Coprocessor prescale. More...

struct	pq_config_t
	powerquad data structure format More...

struct	pq_biquad_param_t
	Struct to save biquad parameters. More...

struct	pq_biquad_state_t
	Struct to save biquad state. More...

struct	pq_biquad_cascade_df2_instance
	Instance structure for the direct form II Biquad cascade filter. More...

union	pq_float_t
	Conversion between integer and float type. More...

Macros
#define	PQ_LN_INF PQ_LN, 1, PQ_TRANS
	Parameter used for vector ln(x)

#define	PQ_INV_INF PQ_INV, 0, PQ_TRANS
	Parameter used for vector 1/x.

#define	PQ_SQRT_INF PQ_SQRT, 0, PQ_TRANS
	Parameter used for vector sqrt(x)

#define	PQ_ISQRT_INF PQ_INVSQRT, 0, PQ_TRANS
	Parameter used for vector 1/sqrt(x)

#define	PQ_ETOX_INF PQ_ETOX, 0, PQ_TRANS
	Parameter used for vector e^x.

#define	PQ_ETONX_INF PQ_ETONX, 0, PQ_TRANS
	Parameter used for vector e^(-x)

#define	PQ_SIN_INF PQ_SIN, 1, PQ_TRIG
	Parameter used for vector sin(x)

#define	PQ_COS_INF PQ_COS, 1, PQ_TRIG
	Parameter used for vector cos(x)

#define	PQ_Initiate_Vector_Func(pSrc, pDst)
	Start 32-bit data vector calculation. More...

#define	PQ_End_Vector_Func() __asm volatile("POP {r2-r7}")
	End vector calculation. More...

#define	PQ_StartVector(PSRC, PDST, LENGTH)
	Start 32-bit data vector calculation. More...

#define	PQ_StartVectorFixed16(PSRC, PDST, LENGTH)
	Start 16-bit data vector calculation. More...

#define	PQ_StartVectorQ15(PSRC, PDST, LENGTH)
	Start Q15-bit data vector calculation. More...

#define	PQ_EndVector() __asm volatile("POP {r3-r10} \n")
	End vector calculation. More...

#define	PQ_Vector8F32(BATCH_OPCODE, DOUBLE_READ_ADDERS, BATCH_MACHINE)
	Float data vector calculation. More...

#define	PQ_Vector8Fixed32(BATCH_OPCODE, DOUBLE_READ_ADDERS, BATCH_MACHINE)
	Fixed 32bits data vector calculation. More...

#define	PQ_Vector8Fixed16(BATCH_OPCODE, DOUBLE_READ_ADDERS, BATCH_MACHINE)
	Fixed 32bits data vector calculation. More...

#define	PQ_Vector8Q15(BATCH_OPCODE, DOUBLE_READ_ADDERS, BATCH_MACHINE)
	Q15 data vector calculation. More...

#define	PQ_DF2_Vector8_FP(middle, last)
	Float data vector biquad direct form II calculation. More...

#define	PQ_DF2_Vector8_FX(middle, last)
	Fixed data vector biquad direct form II calculation. More...

#define	PQ_Vector8BiquadDf2F32()
	Float data vector biquad direct form II calculation. More...

#define	PQ_Vector8BiquadDf2Fixed32()
	Fixed 32-bit data vector biquad direct form II calculation. More...

#define	PQ_Vector8BiquadDf2Fixed16()
	Fixed 16-bit data vector biquad direct form II calculation. More...

#define	PQ_DF2_Cascade_Vector8_FP(middle, last)
	Float data vector direct form II biquad cascade filter. More...

#define	PQ_DF2_Cascade_Vector8_FX(middle, last)
	Fixed data vector direct form II biquad cascade filter. More...

#define	PQ_Vector8BiqaudDf2CascadeF32()
	Float data vector direct form II biquad cascade filter. More...

#define	PQ_Vector8BiqaudDf2CascadeFixed32()
	Fixed 32-bit data vector direct form II biquad cascade filter. More...

#define	PQ_Vector8BiqaudDf2CascadeFixed16()
	Fixed 16-bit data vector direct form II biquad cascade filter. More...

#define	POWERQUAD_MAKE_MATRIX_LEN(mat1Row, mat1Col, mat2Col) (((uint32_t)(mat1Row) << 0U) \| ((uint32_t)(mat1Col) << 8U) \| ((uint32_t)(mat2Col) << 16U))
	Make the length used for matrix functions. More...

#define	PQ_Q31_2_FLOAT(x) (((float)(x)) / 2147483648.0f)
	Convert Q31 to float. More...

#define	PQ_Q15_2_FLOAT(x) (((float)(x)) / 32768.0f)
	Convert Q15 to float. More...

Enumerations
enum	pq_computationengine_t { kPQ_CP_PQ = 0, kPQ_CP_MTX = 1, kPQ_CP_FFT = 2, kPQ_CP_FIR = 3, kPQ_CP_CORDIC = 5 }
	powerquad computation engine More...

enum	pq_format_t { kPQ_16Bit = 0, kPQ_32Bit = 1, kPQ_Float = 2 }
	powerquad data structure format type More...

enum	pq_cordic_iter_t { kPQ_Iteration_8 = 0, kPQ_Iteration_16, kPQ_Iteration_24 }
	CORDIC iteration. More...

Driver version
#define	FSL_POWERQUAD_DRIVER_VERSION (MAKE_VERSION(2, 0, 1))
	Version 2.0.1. More...

POWERQUAD functional Operation
void	PQ_GetDefaultConfig (pq_config_t *config)
	Get default configuration. More...

void	PQ_SetConfig (POWERQUAD_Type base, const pq_config_t config)
	Set configuration with format/prescale. More...

static void	PQ_SetCoprocessorScaler (POWERQUAD_Type base, const pq_prescale_t prescale)
	set coprocessor scaler for coprocessor instructions, this function is used to set output saturation and scaleing for input/output. More...

void	PQ_Init (POWERQUAD_Type *base)
	Initializes the POWERQUAD module. More...

void	PQ_Deinit (POWERQUAD_Type *base)
	De-initializes the POWERQUAD module. More...

void	PQ_SetFormat (POWERQUAD_Type *base, pq_computationengine_t engine, pq_format_t format)
	Set format for non-coprecessor instructions. More...

static void	PQ_WaitDone (POWERQUAD_Type *base)
	Wait for the completion. More...

static void	PQ_LnF32 (float pSrc, float pDst)
	Processing function for the floating-point natural log. More...

static void	PQ_InvF32 (float pSrc, float pDst)
	Processing function for the floating-point reciprocal. More...

static void	PQ_SqrtF32 (float pSrc, float pDst)
	Processing function for the floating-point square-root. More...

static void	PQ_InvSqrtF32 (float pSrc, float pDst)
	Processing function for the floating-point inverse square-root. More...

static void	PQ_EtoxF32 (float pSrc, float pDst)
	Processing function for the floating-point natural exponent. More...

static void	PQ_EtonxF32 (float pSrc, float pDst)
	Processing function for the floating-point natural exponent with negative parameter. More...

static void	PQ_SinF32 (float pSrc, float pDst)
	Processing function for the floating-point sine. More...

static void	PQ_CosF32 (float pSrc, float pDst)
	Processing function for the floating-point cosine. More...

static void	PQ_BiquadF32 (float pSrc, float pDst)
	Processing function for the floating-point biquad. More...

static void	PQ_DivF32 (float x1, float x2, float *pDst)
	Processing function for the floating-point division. More...

static void	PQ_Biquad1F32 (float pSrc, float pDst)
	Processing function for the floating-point biquad. More...

static int32_t	PQ_LnFixed (int32_t val)
	Processing function for the fixed natural log. More...

static int32_t	PQ_InvFixed (int32_t val)
	Processing function for the fixed reciprocal. More...

static uint32_t	PQ_SqrtFixed (uint32_t val)
	Processing function for the fixed square-root. More...

static int32_t	PQ_InvSqrtFixed (int32_t val)
	Processing function for the fixed inverse square-root. More...

static int32_t	PQ_EtoxFixed (int32_t val)
	Processing function for the Fixed natural exponent. More...

static int32_t	PQ_EtonxFixed (int32_t val)
	Processing function for the fixed natural exponent with negative parameter. More...

static int32_t	PQ_SinQ31 (int32_t val)
	Processing function for the fixed sine. More...

static int16_t	PQ_SinQ15 (int16_t val)
	Processing function for the fixed sine. More...

static int32_t	PQ_CosQ31 (int32_t val)
	Processing function for the fixed cosine. More...

static int16_t	PQ_CosQ15 (int16_t val)
	Processing function for the fixed sine. More...

static int32_t	PQ_BiquadFixed (int32_t val)
	Processing function for the fixed biquad. More...

void	PQ_VectorLnF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised natural log. More...

void	PQ_VectorInvF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised reciprocal. More...

void	PQ_VectorSqrtF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised square-root. More...

void	PQ_VectorInvSqrtF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised inverse square-root. More...

void	PQ_VectorEtoxF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised natural exponent. More...

void	PQ_VectorEtonxF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised natural exponent with negative parameter. More...

void	PQ_VectorSinF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised sine. More...

void	PQ_VectorCosF32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised cosine. More...

void	PQ_VectorLnFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the Q31 vectorised natural log. More...

void	PQ_VectorInvFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the Q31 vectorised reciprocal. More...

void	PQ_VectorSqrtFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised square-root. More...

void	PQ_VectorInvSqrtFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised inverse square-root. More...

void	PQ_VectorEtoxFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised natural exponent. More...

void	PQ_VectorEtonxFixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised natural exponent with negative parameter. More...

void	PQ_VectorSinQ15 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the Q15 vectorised sine. More...

void	PQ_VectorCosQ15 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the Q15 vectorised cosine. More...

void	PQ_VectorSinQ31 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the Q31 vectorised sine. More...

void	PQ_VectorCosQ31 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the Q31 vectorised cosine. More...

void	PQ_VectorLnFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised natural log. More...

void	PQ_VectorInvFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised reciprocal. More...

void	PQ_VectorSqrtFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised square-root. More...

void	PQ_VectorInvSqrtFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised inverse square-root. More...

void	PQ_VectorEtoxFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised natural exponent. More...

void	PQ_VectorEtonxFixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised natural exponent with negative parameter. More...

void	PQ_VectorBiqaudDf2F32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised biquad direct form II. More...

void	PQ_VectorBiqaudDf2Fixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised biquad direct form II. More...

void	PQ_VectorBiqaudDf2Fixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised biquad direct form II. More...

void	PQ_VectorBiqaudCascadeDf2F32 (float pSrc, float pDst, int32_t length)
	Processing function for the floating-point vectorised biquad direct form II. More...

void	PQ_VectorBiqaudCascadeDf2Fixed32 (int32_t pSrc, int32_t pDst, int32_t length)
	Processing function for the 32-bit integer vectorised biquad direct form II. More...

void	PQ_VectorBiqaudCascadeDf2Fixed16 (int16_t pSrc, int16_t pDst, int32_t length)
	Processing function for the 16-bit integer vectorised biquad direct form II. More...

int32_t	PQ_ArctanFixed (POWERQUAD_Type *base, int32_t x, int32_t y, pq_cordic_iter_t iteration)
	Processing function for the fixed inverse trigonometric. More...

int32_t	PQ_ArctanhFixed (POWERQUAD_Type *base, int32_t x, int32_t y, pq_cordic_iter_t iteration)
	Processing function for the fixed inverse trigonometric. More...

static int32_t	PQ_Biquad1Fixed (int32_t val)
	Processing function for the fixed biquad. More...

void	PQ_TransformCFFT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the complex FFT. More...

void	PQ_TransformRFFT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the real FFT. More...

void	PQ_TransformIFFT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the inverse complex FFT. More...

void	PQ_TransformCDCT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the complex DCT. More...

void	PQ_TransformRDCT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the real DCT. More...

void	PQ_TransformIDCT (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the inverse complex DCT. More...

void	PQ_BiquadBackUpInternalState (POWERQUAD_Type base, int32_t biquad_num, pq_biquad_state_t state)
	Processing function for backup biquad context. More...

void	PQ_BiquadRestoreInternalState (POWERQUAD_Type base, int32_t biquad_num, pq_biquad_state_t state)
	Processing function for restore biquad context. More...

void	PQ_BiquadCascadeDf2Init (pq_biquad_cascade_df2_instance S, uint8_t numStages, pq_biquad_state_t pState)
	Initialization function for the direct form II Biquad cascade filter. More...

void	PQ_BiquadCascadeDf2F32 (const pq_biquad_cascade_df2_instance S, float pSrc, float *pDst, uint32_t blockSize)
	Processing function for the floating-point direct form II Biquad cascade filter. More...

void	PQ_BiquadCascadeDf2Fixed32 (const pq_biquad_cascade_df2_instance S, int32_t pSrc, int32_t *pDst, uint32_t blockSize)
	Processing function for the Q31 direct form II Biquad cascade filter. More...

void	PQ_BiquadCascadeDf2Fixed16 (const pq_biquad_cascade_df2_instance S, int16_t pSrc, int16_t *pDst, uint32_t blockSize)
	Processing function for the Q15 direct form II Biquad cascade filter. More...

void	PQ_FIR (POWERQUAD_Type base, void pAData, int32_t ALength, void pBData, int32_t BLength, void pResult, uint32_t opType)
	Processing function for the FIR. More...

void	PQ_FIRIncrement (POWERQUAD_Type *base, int32_t ALength, int32_t BLength, int32_t xOffset)
	Processing function for the incremental FIR. More...

void	PQ_MatrixAddition (POWERQUAD_Type base, uint32_t length, void pAData, void pBData, void pResult)
	Processing function for the matrix addition. More...

void	PQ_MatrixSubtraction (POWERQUAD_Type base, uint32_t length, void pAData, void pBData, void pResult)
	Processing function for the matrix subtraction. More...

void	PQ_MatrixMultiplication (POWERQUAD_Type base, uint32_t length, void pAData, void pBData, void pResult)
	Processing function for the matrix multiplication. More...

void	PQ_MatrixProduct (POWERQUAD_Type base, uint32_t length, void pAData, void pBData, void pResult)
	Processing function for the matrix product. More...

void	PQ_VectorDotProduct (POWERQUAD_Type base, uint32_t length, void pAData, void pBData, void pResult)
	Processing function for the vector dot product. More...

void	PQ_MatrixInversion (POWERQUAD_Type base, uint32_t length, void pData, void pTmpData, void pResult)
	Processing function for the matrix inverse. More...

void	PQ_MatrixTranspose (POWERQUAD_Type base, uint32_t length, void pData, void *pResult)
	Processing function for the matrix transpose. More...

void	PQ_MatrixScale (POWERQUAD_Type base, uint32_t length, float misc, void pData, void *pResult)
	Processing function for the matrix scale. More...

Data Structure Documentation

struct pq_prescale_t

Data Fields
int8_t	inputPrescale
	Input prescale. More...

int8_t	outputPrescale
	Output prescale. More...

int8_t	outputSaturate
	Output saturate at n bits, for example 0x11 is 8 bit space, the value will be truncated at +127 or -128. More...

Field Documentation

int8_t pq_prescale_t::inputPrescale

int8_t pq_prescale_t::outputPrescale

int8_t pq_prescale_t::outputSaturate

struct pq_config_t

Data Fields
pq_format_t	inputAFormat
	Input A format. More...

int8_t	inputAPrescale
	Input A prescale, for example 1.5 can be 1.5*2^n if you scale by 'shifting' ('scaling' by a factor of n). More...

pq_format_t	inputBFormat
	Input B format. More...

int8_t	inputBPrescale
	Input B prescale. More...

pq_format_t	outputFormat
	Out format. More...

int8_t	outputPrescale
	Out prescale. More...

pq_format_t	tmpFormat
	Temp format. More...

int8_t	tmpPrescale
	Temp prescale. More...

pq_format_t	machineFormat
	Machine format. More...

uint32_t *	tmpBase
	Tmp base address. More...

Field Documentation

pq_format_t pq_config_t::inputAFormat

int8_t pq_config_t::inputAPrescale

pq_format_t pq_config_t::inputBFormat

int8_t pq_config_t::inputBPrescale

pq_format_t pq_config_t::outputFormat

int8_t pq_config_t::outputPrescale

pq_format_t pq_config_t::tmpFormat

int8_t pq_config_t::tmpPrescale

pq_format_t pq_config_t::machineFormat

uint32_t* pq_config_t::tmpBase

struct pq_biquad_param_t

Data Fields
float	v_n_1
	v[n-1], set to 0 when initialization. More...

float	v_n
	v[n], set to 0 when initialization. More...

float	a_1
	a[1]

float	a_2
	a[2]

float	b_0
	b[0]

float	b_1
	b[1]

float	b_2
	b[2]

Field Documentation

float pq_biquad_param_t::v_n_1

float pq_biquad_param_t::v_n

struct pq_biquad_state_t

Data Fields
pq_biquad_param_t	param
	Filter parameter. More...

uint32_t	compreg
	Internal register, set to 0 when initialization. More...

Field Documentation

pq_biquad_param_t pq_biquad_state_t::param

uint32_t pq_biquad_state_t::compreg

struct pq_biquad_cascade_df2_instance

Data Fields
uint8_t	numStages

pq_biquad_state_t *	pState

Field Documentation

uint8_t pq_biquad_cascade_df2_instance::numStages

Number of 2nd order stages in the filter.

pq_biquad_state_t* pq_biquad_cascade_df2_instance::pState

Points to the array of state coefficients.

union pq_float_t

Data Fields
float	floatX
	Float type. More...

uint32_t	integerX
	Iterger type. More...

Field Documentation

float pq_float_t::floatX

uint32_t pq_float_t::integerX

Macro Definition Documentation

#define FSL_POWERQUAD_DRIVER_VERSION (MAKE_VERSION(2, 0, 1))

#define PQ_Initiate_Vector_Func	(	pSrc,
		pDst
	)

Value:

__asm volatile(                                      \
        "MOV r0, %[psrc]         \n"                     \
        "MOV r1, %[pdst]         \n"                     \
        "PUSH {r2-r7}            \n"                     \
        "LDRD r2,r3,[r0],#8      \n" ::[psrc] "r"(pSrc), \
        [pdst] "r"(pDst)                                 \
        : "r0", "r1")

Start the vector calculation, the input data could be float, int32_t or Q31.

Parameters

PSRC	Pointer to the source data.
PDST	Pointer to the destination data.

#define PQ_End_Vector_Func ( ) __asm volatile("POP {r2-r7}")

This function should be called after vector calculation.

#define PQ_StartVector	(	PSRC,
		PDST,
		LENGTH
	)

Value:

__asm volatile(                                      \
        "MOV r0, %[psrc]         \n"                     \
        "MOV r1, %[pdst]         \n"                     \
        "MOV r2, %[length]       \n"                     \
        "PUSH {r3-r10}           \n"                     \
        "MOV r3, #0              \n"                     \
        "MOV r10, #0             \n"                     \
        "LDRD r4,r5,[r0],#8      \n" ::[psrc] "r"(PSRC), \
        [pdst] "r"(PDST), [length] "r"(LENGTH)           \
        : "r0", "r1", "r2")

Start the vector calculation, the input data could be float, int32_t or Q31.

Parameters

PSRC	Pointer to the source data.
PDST	Pointer to the destination data.
LENGTH	Number of the data, must be multiple of 8.

#define PQ_StartVectorFixed16	(	PSRC,
		PDST,
		LENGTH
	)

Value:

__asm volatile(                                       \
        "MOV r0, %[psrc]          \n"                     \
        "MOV r1, %[pdst]          \n"                     \
        "MOV r2, %[length]        \n"                     \
        "PUSH {r3-r10}            \n"                     \
        "MOV r3, #0               \n"                     \
        "LDRSH r4,[r0],#2         \n"                     \
        "LDRSH r5,[r0],#2         \n" ::[psrc] "r"(PSRC), \
        [pdst] "r"(PDST), [length] "r"(LENGTH)            \
        : "r0", "r1", "r2")

Start the vector calculation, the input data could be int16_t. This function should be use with PQ_Vector8Fixed16.

Parameters

PSRC	Pointer to the source data.
PDST	Pointer to the destination data.
LENGTH	Number of the data, must be multiple of 8.

#define PQ_StartVectorQ15	(	PSRC,
		PDST,
		LENGTH
	)

Value:

__asm volatile(                                       \
        "MOV r0, %[psrc]          \n"                     \
        "MOV r1, %[pdst]          \n"                     \
        "MOV r2, %[length]        \n"                     \
        "PUSH {r3-r10}            \n"                     \
        "MOV r3, #0               \n"                     \
        "LDR r5,[r0],#4           \n"                     \
        "LSL r4,r5,#16            \n"                     \
        "BFC r5,#0,#16            \n" ::[psrc] "r"(PSRC), \
        [pdst] "r"(PDST), [length] "r"(LENGTH)            \
        : "r0", "r1", "r2")

Start the vector calculation, the input data could be Q15. This function should be use with PQ_Vector8Q15. This function is dedicate for SinQ15/CosQ15 vector calculation. Because PowerQuad only supports Q31 Sin/Cos fixed function, so the input Q15 data is left shift 16 bits first, after Q31 calculation, the output data is right shift 16 bits.

Parameters

PSRC	Pointer to the source data.
PDST	Pointer to the destination data.
LENGTH	Number of the data, must be multiple of 8.

#define PQ_EndVector ( ) __asm volatile("POP {r3-r10} \n")

This function should be called after vector calculation.

#define PQ_Vector8F32	(	BATCH_OPCODE,
		DOUBLE_READ_ADDERS,
		BATCH_MACHINE
	)

Float data vector calculation, the input data should be float. The parameter could be PQ_LN_INF, PQ_INV_INF, PQ_SQRT_INF, PQ_ISQRT_INF, PQ_ETOX_INF, PQ_ETONX_INF. For example, to calculate sqrt of a vector, use like this:

#define VECTOR_LEN 8
float input[VECTOR_LEN] = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0};
float output[VECTOR_LEN];
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8F32(PQ_SQRT_INF);
PQ_EndVector();

#define PQ_Vector8Fixed32	(	BATCH_OPCODE,
		DOUBLE_READ_ADDERS,
		BATCH_MACHINE
	)

Float data vector calculation, the input data should be 32-bit integer. The parameter could be PQ_LN_INF, PQ_INV_INF, PQ_SQRT_INF, PQ_ISQRT_INF, PQ_ETOX_INF, PQ_ETONX_INF. PQ_SIN_INF, PQ_COS_INF. When this function is used for sin/cos calculation, the input data should be in the format Q1.31. For example, to calculate sqrt of a vector, use like this:

#define VECTOR_LEN 8
int32_t input[VECTOR_LEN] = {1, 4, 9, 16, 25, 36, 49, 64};
int32_t output[VECTOR_LEN];
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8F32(PQ_SQRT_INF);
PQ_EndVector();

#define PQ_Vector8Fixed16	(	BATCH_OPCODE,
		DOUBLE_READ_ADDERS,
		BATCH_MACHINE
	)

Float data vector calculation, the input data should be 16-bit integer. The parameter could be PQ_LN_INF, PQ_INV_INF, PQ_SQRT_INF, PQ_ISQRT_INF, PQ_ETOX_INF, PQ_ETONX_INF. For example, to calculate sqrt of a vector, use like this:

#define VECTOR_LEN 8
int16_t input[VECTOR_LEN] = {1, 4, 9, 16, 25, 36, 49, 64};
int16_t output[VECTOR_LEN];
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8F32(PQ_SQRT_INF);
PQ_EndVector();

#define PQ_Vector8Q15	(	BATCH_OPCODE,
		DOUBLE_READ_ADDERS,
		BATCH_MACHINE
	)

Q15 data vector calculation, this function should only be used for sin/cos Q15 calculation, and the coprocessor output prescaler must be set to 31 before this function. This function loads Q15 data and left shift 16 bits, calculate and right shift 16 bits, then stores to the output array. The input range -1 to 1 means -pi to pi. For example, to calculate sin of a vector, use like this:

#define VECTOR_LEN 8
int16_t input[VECTOR_LEN] = {...}
int16_t output[VECTOR_LEN];
const pq_prescale_t prescale =
{
    .inputPrescale = 0,
    .outputPrescale = 31,
    .outputSaturate = 0
};
PQ_SetCoprocessorScaler(POWERQUAD, const pq_prescale_t *prescale);
PQ_StartVectorQ15(pSrc, pDst, length);
PQ_Vector8Q15(PQ_SQRT_INF);
PQ_EndVector();

#define PQ_DF2_Vector8_FP	(	middle,
		last
	)

Biquad filter, the input and output data are float data. Biquad side 0 is used. Example:

#define VECTOR_LEN 16
float input[VECTOR_LEN] = {1024.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
float output[VECTOR_LEN];
pq_biquad_state_t state =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state);
PQ_Initiate_Vector_Func(pSrc,pDst);
PQ_DF2_Vector8_FP(false,false);
PQ_DF2_Vector8_FP(true,true);
PQ_End_Vector_Func();

#define PQ_DF2_Vector8_FX	(	middle,
		last
	)

Biquad filter, the input and output data are fixed data. Biquad side 0 is used. Example:

#define VECTOR_LEN 16
int32_t input[VECTOR_LEN] = {1024, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
int32_t output[VECTOR_LEN];
pq_biquad_state_t state =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state);
PQ_Initiate_Vector_Func(pSrc,pDst);
PQ_DF2_Vector8_FX(false,false);
PQ_DF2_Vector8_FX(true,true);
PQ_End_Vector_Func();

#define PQ_Vector8BiquadDf2F32 ( )

Biquad filter, the input and output data are float data. Biquad side 0 is used. Example:

#define VECTOR_LEN 8
float input[VECTOR_LEN] = {1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0};
float output[VECTOR_LEN];
pq_biquad_state_t state =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiquadDf2F32();
PQ_EndVector();

#define PQ_Vector8BiquadDf2Fixed32 ( )

Biquad filter, the input and output data are Q31 or 32-bit integer. Biquad side 0 is used. Example:

#define VECTOR_LEN 8
int32_t input[VECTOR_LEN] = {1, 2, 3, 4, 5, 6, 7, 8};
int32_t output[VECTOR_LEN];
pq_biquad_state_t state =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiquadDf2Fixed32();
PQ_EndVector();

#define PQ_Vector8BiquadDf2Fixed16 ( )

Biquad filter, the input and output data are Q15 or 16-bit integer. Biquad side 0 is used. Example:

#define VECTOR_LEN 8
int16_t input[VECTOR_LEN] = {1, 2, 3, 4, 5, 6, 7, 8};
int16_t output[VECTOR_LEN];
pq_biquad_state_t state =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiquadDf2Fixed16();
PQ_EndVector();

#define PQ_DF2_Cascade_Vector8_FP	(	middle,
		last
	)

The input and output data are float data. The data flow is input -> biquad side 1 -> biquad side 0 -> output.

#define VECTOR_LEN 16
float input[VECTOR_LEN] = {1024.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
float output[VECTOR_LEN];
pq_biquad_state_t state0 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
pq_biquad_state_t state1 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state0);
PQ_BiquadRestoreInternalState(POWERQUAD, 1, &state1);
PQ_Initiate_Vector_Func(pSrc, pDst);
PQ_DF2_Cascade_Vector8_FP(false, false);
PQ_DF2_Cascade_Vector8_FP(true, true);
PQ_End_Vector_Func();

#define PQ_DF2_Cascade_Vector8_FX	(	middle,
		last
	)

The input and output data are fixed data. The data flow is input -> biquad side 1 -> biquad side 0 -> output.

#define VECTOR_LEN 16
int32_t input[VECTOR_LEN] = {1024.0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
int32_t output[VECTOR_LEN];
pq_biquad_state_t state0 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
pq_biquad_state_t state1 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state0);
PQ_BiquadRestoreInternalState(POWERQUAD, 1, &state1);
PQ_Initiate_Vector_Func(pSrc, pDst);
PQ_DF2_Cascade_Vector8_FX(false, false);
PQ_DF2_Cascade_Vector8_FX(true, true);
PQ_End_Vector_Func();

#define PQ_Vector8BiqaudDf2CascadeF32 ( )

The input and output data are float data. The data flow is input -> biquad side 1 -> biquad side 0 -> output.

#define VECTOR_LEN 8
float input[VECTOR_LEN] = {1, 2, 3, 4, 5, 6, 7, 8};
float output[VECTOR_LEN];
pq_biquad_state_t state0 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
pq_biquad_state_t state1 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state0);
PQ_BiquadRestoreInternalState(POWERQUAD, 1, &state1);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiqaudDf2CascadeF32();
PQ_EndVector();

#define PQ_Vector8BiqaudDf2CascadeFixed32 ( )

The input and output data are fixed 32-bit data. The data flow is input -> biquad side 1 -> biquad side 0 -> output.

#define VECTOR_LEN 8
int32_t input[VECTOR_LEN] = {1, 2, 3, 4, 5, 6, 7, 8};
int32_t output[VECTOR_LEN];
pq_biquad_state_t state0 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
pq_biquad_state_t state1 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state0);
PQ_BiquadRestoreInternalState(POWERQUAD, 1, &state1);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiqaudDf2CascadeFixed32();
PQ_EndVector();

#define PQ_Vector8BiqaudDf2CascadeFixed16 ( )

The input and output data are fixed 16-bit data. The data flow is input -> biquad side 1 -> biquad side 0 -> output.

#define VECTOR_LEN 8
int32_t input[VECTOR_LEN] = {1, 2, 3, 4, 5, 6, 7, 8};
int32_t output[VECTOR_LEN];
pq_biquad_state_t state0 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
pq_biquad_state_t state1 =
{
     .param =
     {
         .a_1 = xxx,
         .a_2 = xxx,
         .b_0 = xxx,
         .b_1 = xxx,
         .b_2 = xxx,
     },
};
PQ_BiquadRestoreInternalState(POWERQUAD, 0, &state0);
PQ_BiquadRestoreInternalState(POWERQUAD, 1, &state1);
PQ_StartVector(input, output, VECTOR_LEN);
PQ_Vector8BiqaudDf2CascadeFixed16();
PQ_EndVector();

#define POWERQUAD_MAKE_MATRIX_LEN	(	mat1Row,
		mat1Col,
		mat2Col
	)	(((uint32_t)(mat1Row) << 0U) \| ((uint32_t)(mat1Col) << 8U) \| ((uint32_t)(mat2Col) << 16U))

#define PQ_Q31_2_FLOAT ( x ) (((float)(x)) / 2147483648.0f)

#define PQ_Q15_2_FLOAT ( x ) (((float)(x)) / 32768.0f)

Enumeration Type Documentation

enum pq_computationengine_t

Enumerator
kPQ_CP_PQ	Math engine.
kPQ_CP_MTX	Matrix engine.
kPQ_CP_FFT	FFT engine.
kPQ_CP_FIR	FIR engine.
kPQ_CP_CORDIC	CORDIC engine.

enum pq_format_t

Enumerator
kPQ_16Bit	Int16 Fixed point.
kPQ_32Bit	Int32 Fixed point.
kPQ_Float	Float point.

enum pq_cordic_iter_t

Enumerator
kPQ_Iteration_8	Iterate 8 times.
kPQ_Iteration_16	Iterate 16 times.
kPQ_Iteration_24	Iterate 24 times.

Function Documentation

void PQ_GetDefaultConfig ( pq_config_t * config )

This function initializes the POWERQUAD configuration structure to a default value. FORMAT register field definitions Bits[15:8] scaler (for scaled 'q31' formats) Bits[5:4] external format. 00b=q15, 01b=q31, 10b=float Bits[1:0] internal format. 00b=q15, 01b=q31, 10b=float POWERQUAD->INAFORMAT = (config->inputAPrescale << 8) | (config->inputAFormat << 4) | config->machineFormat

For all Powerquad operations internal format must be float (with the only exception being the FFT related functions, ie FFT/IFFT/DCT/IDCT which must be set to q31). The default values are: config->inputAFormat = kPQ_Float; config->inputAPrescale = 0; config->inputBFormat = kPQ_Float; config->inputBPrescale = 0; config->outputFormat = kPQ_Float; config->outputPrescale = 0; config->tmpFormat = kPQ_Float; config->tmpPrescale = 0; config->machineFormat = kPQ_Float; config->tmpBase = 0xE0000000;

Parameters

config Pointer to "pq_config_t" structure.

void PQ_SetConfig	(	POWERQUAD_Type *	base,
		const pq_config_t *	config
	)

Parameters

base	POWERQUAD peripheral base address
config	Pointer to "pq_config_t" structure.

static void PQ_SetCoprocessorScaler	(	POWERQUAD_Type *	base,
		const pq_prescale_t *	prescale
	)

inlinestatic

Parameters

base	POWERQUAD peripheral base address
prescale	Pointer to "pq_prescale_t" structure.

void PQ_Init ( POWERQUAD_Type * base )

Parameters

base	POWERQUAD peripheral base address.

void PQ_Deinit ( POWERQUAD_Type * base )

Parameters

base	POWERQUAD peripheral base address.

void PQ_SetFormat	(	POWERQUAD_Type *	base,
		pq_computationengine_t	engine,
		pq_format_t	format
	)

Parameters

base	POWERQUAD peripheral base address
engine	Computation engine
format	Data format

static void PQ_WaitDone ( POWERQUAD_Type * base )

inlinestatic

Parameters

base	POWERQUAD peripheral base address

static void PQ_LnF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is (0 +INFINITY).
*pDst	points to the block of output data

static void PQ_InvF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is non-zero.
*pDst	points to the block of output data

static void PQ_SqrtF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is [0 +INFINITY).
*pDst	points to the block of output data

static void PQ_InvSqrtF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is (0 +INFINITY).
*pDst	points to the block of output data

static void PQ_EtoxF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is (-INFINITY +INFINITY).
*pDst	points to the block of output data

static void PQ_EtonxF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The range of the input value is (-INFINITY +INFINITY).
*pDst	points to the block of output data

static void PQ_SinF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The input value is in radians, the range is (-INFINITY +INFINITY).
*pDst	points to the block of output data

static void PQ_CosF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data. The input value is in radians, the range is (-INFINITY +INFINITY).
*pDst	points to the block of output data

static void PQ_BiquadF32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data

static void PQ_DivF32	(	float *	x1,
		float *	x2,
		float *	pDst
	)

inlinestatic

Get x1 / x2.

Parameters

x1	x1
x2	x2
*pDst	points to the block of output data

static void PQ_Biquad1F32	(	float *	pSrc,
		float *	pDst
	)

inlinestatic

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data

static int32_t PQ_LnFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is (0 +INFINITY).

Returns: returns ln(val).

static int32_t PQ_InvFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is non-zero.

Returns: returns inv(val).

static uint32_t PQ_SqrtFixed ( uint32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is [0 +INFINITY).

Returns: returns sqrt(val).

static int32_t PQ_InvSqrtFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is (0 +INFINITY).

Returns: returns 1/sqrt(val).

static int32_t PQ_EtoxFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is (-INFINITY +INFINITY).

Returns: returns etox^(val).

static int32_t PQ_EtonxFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The range of the input value is (-INFINITY +INFINITY).

Returns: returns etonx^(val).

static int32_t PQ_SinQ31 ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The input value is [-1, 1] in Q31 format, which means [-pi, pi].

Returns: returns sin(val).

static int16_t PQ_SinQ15 ( int16_t val )

inlinestatic

Parameters

val	value to be calculated. The input value is [-1, 1] in Q15 format, which means [-pi, pi].

Returns: returns sin(val).

static int32_t PQ_CosQ31 ( int32_t val )

inlinestatic

Parameters

val	value to be calculated. The input value is [-1, 1] in Q31 format, which means [-pi, pi].

Returns: returns cos(val).

static int16_t PQ_CosQ15 ( int16_t val )

inlinestatic

Parameters

val	value to be calculated. The input value is [-1, 1] in Q15 format, which means [-pi, pi].

Returns: returns sin(val).

static int32_t PQ_BiquadFixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated

Returns: returns biquad(val).

void PQ_VectorLnF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSqrtF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvSqrtF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtoxF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtonxF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSinF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorCosF32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorLnFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSqrtFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvSqrtFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtoxFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtonxFixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSinQ15	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorCosQ15	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSinQ31	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorCosQ31	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorLnFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorSqrtFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorInvSqrtFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtoxFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorEtonxFixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block of input data.

void PQ_VectorBiqaudDf2F32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
length	the block size of input data.

void PQ_VectorBiqaudDf2Fixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
blocksSize	the block size of input data

void PQ_VectorBiqaudDf2Fixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
blocksSize	the block size of input data

void PQ_VectorBiqaudCascadeDf2F32	(	float *	pSrc,
		float *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
blocksSize	the block size of input data

void PQ_VectorBiqaudCascadeDf2Fixed32	(	int32_t *	pSrc,
		int32_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
blocksSize	the block size of input data

void PQ_VectorBiqaudCascadeDf2Fixed16	(	int16_t *	pSrc,
		int16_t *	pDst,
		int32_t	length
	)

Parameters

*pSrc	points to the block of input data
*pDst	points to the block of output data
blocksSize	the block size of input data

int32_t PQ_ArctanFixed	(	POWERQUAD_Type *	base,
		int32_t	x,
		int32_t	y,
		pq_cordic_iter_t	iteration
	)

Parameters

base	POWERQUAD peripheral base address
x	value of opposite
y	value of adjacent
iteration	iteration times

Returns: The return value is in the range of -2^27 to 2^27, which means -pi to pi.

Note: The sum of x and y should not exceed the range of int32_t.; Larger input number gets higher output accuracy, for example the arctan(0.5), the result of PQ_ArctanFixed(POWERQUAD, 100000, 200000, kPQ_Iteration_24) is more accurate than PQ_ArctanFixed(POWERQUAD, 1, 2, kPQ_Iteration_24).

int32_t PQ_ArctanhFixed	(	POWERQUAD_Type *	base,
		int32_t	x,
		int32_t	y,
		pq_cordic_iter_t	iteration
	)

Parameters

base	POWERQUAD peripheral base address
x	value of opposite
y	value of adjacent
iteration	iteration times

Returns: The return value is in the range of -2^27 to 2^27, which means -1 to 1.

Note: The sum of x and y should not exceed the range of int32_t.; Larger input number gets higher output accuracy, for example the arctanh(0.5), the result of PQ_ArctanhFixed(POWERQUAD, 100000, 200000, kPQ_Iteration_24) is more accurate than PQ_ArctanhFixed(POWERQUAD, 1, 2, kPQ_Iteration_24).

static int32_t PQ_Biquad1Fixed ( int32_t val )

inlinestatic

Parameters

val	value to be calculated

Returns: returns biquad(val).

void PQ_TransformCFFT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_TransformRFFT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_TransformIFFT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_TransformCDCT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_TransformRDCT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_TransformIDCT	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	number of input samples
pData	input data
pResult	output data.

void PQ_BiquadBackUpInternalState	(	POWERQUAD_Type *	base,
		int32_t	biquad_num,
		pq_biquad_state_t *	state
	)

Parameters

base	POWERQUAD peripheral base address
biquad_num	biquad side
state	point to states.

void PQ_BiquadRestoreInternalState	(	POWERQUAD_Type *	base,
		int32_t	biquad_num,
		pq_biquad_state_t *	state
	)

Parameters

base	POWERQUAD peripheral base address
biquad_num	biquad side
state	point to states.

void PQ_BiquadCascadeDf2Init	(	pq_biquad_cascade_df2_instance *	S,
		uint8_t	numStages,
		pq_biquad_state_t *	pState
	)

Parameters

[in,out]	*S	points to an instance of the filter data structure.
[in]	numStages	number of 2nd order stages in the filter.
[in]	*pState	points to the state buffer.

void PQ_BiquadCascadeDf2F32	(	const pq_biquad_cascade_df2_instance *	S,
		float *	pSrc,
		float *	pDst,
		uint32_t	blockSize
	)

Parameters

[in]	*S	points to an instance of the filter data structure.
[in]	*pSrc	points to the block of input data.
[out]	*pDst	points to the block of output data
[in]	blockSize	number of samples to process.

void PQ_BiquadCascadeDf2Fixed32	(	const pq_biquad_cascade_df2_instance *	S,
		int32_t *	pSrc,
		int32_t *	pDst,
		uint32_t	blockSize
	)

Parameters

[in]	*S	points to an instance of the filter data structure.
[in]	*pSrc	points to the block of input data.
[out]	*pDst	points to the block of output data
[in]	blockSize	number of samples to process.

void PQ_BiquadCascadeDf2Fixed16	(	const pq_biquad_cascade_df2_instance *	S,
		int16_t *	pSrc,
		int16_t *	pDst,
		uint32_t	blockSize
	)

Parameters

[in]	*S	points to an instance of the filter data structure.
[in]	*pSrc	points to the block of input data.
[out]	*pDst	points to the block of output data
[in]	blockSize	number of samples to process.

void PQ_FIR	(	POWERQUAD_Type *	base,
		void *	pAData,
		int32_t	ALength,
		void *	pBData,
		int32_t	BLength,
		void *	pResult,
		uint32_t	opType
	)

Parameters

base	POWERQUAD peripheral base address
pAData	the first input sequence
ALength	number of the first input sequence
pBData	the second input sequence
BLength	number of the second input sequence
pResult	array for the output data
opType	operation type, could be PQ_FIR_FIR, PQ_FIR_CONVOLUTION, PQ_FIR_CORRELATION.

void PQ_FIRIncrement	(	POWERQUAD_Type *	base,
		int32_t	ALength,
		int32_t	BLength,
		int32_t	xOffset
	)

   This function can be used after pq_fir() for incremental FIR
   operation when new x data are available

Parameters

base	POWERQUAD peripheral base address
ALength	number of input samples
BLength	number of taps
xoffset	offset for number of input samples

void PQ_MatrixAddition	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pAData,
		void *	pBData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pAData	input matrix A
pBData	input matrix B
pResult	array for the output data.

void PQ_MatrixSubtraction	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pAData,
		void *	pBData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pAData	input matrix A
pBData	input matrix B
pResult	array for the output data.

void PQ_MatrixMultiplication	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pAData,
		void *	pBData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pAData	input matrix A
pBData	input matrix B
pResult	array for the output data.

void PQ_MatrixProduct	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pAData,
		void *	pBData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pAData	input matrix A
pBData	input matrix B
pResult	array for the output data.

void PQ_VectorDotProduct	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pAData,
		void *	pBData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	length of vector
pAData	input vector A
pBData	input vector B
pResult	array for the output data.

void PQ_MatrixInversion	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pTmpData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pData	input matrix
pTmpData	input temporary matrix, pTmpData length not less than pData lenght and 1024 words is sufficient for the largest supported matrix.
pResult	array for the output data, round down for fixed point.

void PQ_MatrixTranspose	(	POWERQUAD_Type *	base,
		uint32_t	length,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
pData	input matrix
pResult	array for the output data.

void PQ_MatrixScale	(	POWERQUAD_Type *	base,
		uint32_t	length,
		float	misc,
		void *	pData,
		void *	pResult
	)

Parameters

base	POWERQUAD peripheral base address
length	rows and cols for matrix. LENGTH register configuration: LENGTH[23:16] = M2 cols LENGTH[15:8] = M1 cols LENGTH[7:0] = M1 rows This could be constructed using macro POWERQUAD_MAKE_MATRIX_LEN.
misc	scaling parameters
pData	input matrix
pResult	array for the output data.

Overview

Function groups

POWERQUAD functional Operation

Data Structures

Macros

Enumerations

Driver version

POWERQUAD functional Operation

Data Structure Documentation

Data Fields

Field Documentation

Data Fields

Field Documentation

Data Fields

Field Documentation

Data Fields

Field Documentation

Data Fields

Field Documentation

Data Fields

Field Documentation

Macro Definition Documentation

Enumeration Type Documentation

Function Documentation